• How does PG_Bulkload compare to native Copy?
• How do we improve PostgreSQL throughput while using very little memory?
• How fast can it PostgreSQL load 15 Billion rows of data?

Join us for a webinar in which Matt Goldensohn, Director of WHIPTAIL Labs, shares the results, reviews what was learned in the process and discusses the impact for DBAs and their business.

The post Webinar: High Performance Bulk Loading with PostgreSQL appeared first on WHIPTAIL.

Video Transcript:

Hi, my name is Max Riggsbee. I’m the Vice President of Products and Marketing here at WHIPTAIL.

Today I wanted to share with you some really exciting news. We’ve launched what we call WHIPTAIL Labs. What is WHIPTAIL Labs? It’s an idea that was born out of a series of conversations last year. We started asking ourselves, what are the real advantages that customers can leverage from flash technology. Of course, on the surface this seems quite easy—you take an application and you move it to flash and magically it should just go faster. As we thought about this more and more, what we began to realize is nearly every application has been designed with hard disk drives in mind. What that really means is that many of them have spent quite a bit of time avoiding engaging in IO. So WHIPTAIL Labs was born.

What do we do in the Labs? We rethink applications specifically for flash. What does this mean? This means we take an application and we ask ourselves, can we disable cache, can we not use certain features, how do we have multiple applications working together or, in some cases, have a single area of information that multiple users can go after even if they wish to do it in different ways. The output from WHIPTAIL Labs will be real Lab Notes, and this is something that we have already begun to share. We’ve created application groups specifically around flash.

More importantly, the Labs only work with real data: real databases, real virtual desktops, real virtual machines, real applications. We do this because the Labs thinks the way a customer does. We think about how an application may load data, for example. Then we think, how can we load it faster, how can we load more of it faster. Then we go to work. We begin to change parameters around the operating system. We start seeking for features that are seldom used, say, in a database engine. Or we turn things off. That’s right, sometimes we turn off things we’ve all been using for years. Even to our amazement, the results sometimes are somewhat staggering.

I invite you, along with everyone else, to really join in with the work that we’re going to do in the Labs. You’ll see through some of the boards that we’ve created, we’re looking for your feedback. More importantly, we’re looking to give you a window into the future, a window into the way things will work in days to come as you begin to move your applications to flash.

The post Video: About WHIPTAIL Labs appeared first on WHIPTAIL.

Whippany, N.J. – June 18, 2013–WHIPTAIL the data storage-industry innovator powering faster applications and more energy efficient computing for today’s global businesses via flash storage, today announced the appointment of Erik Hardy as executive vice president of worldwide sales. Reporting to CEO Dan Crain, Hardy will lead WHIPTAIL’s Field Operations team in driving global expansion of its powerful flash storage solutions.

“Adding Erik gives us another global-scale thinker as we continue to expand our footprint across all market segments around the world,” said Dan Crain, CEO at WHIPTAIL. “Erik’s unique balance and experience of selling innovative, emerging technologies in storage, networking, software, and services, plus his track record of successful account development makes him a great addition to our team.”

Hardy brings more than 21 years of building and managing world-class sales teams to WHIPTAIL, having spent the majority of his career at EMC, where he served as regional vice president of North American global accounts. After EMC, Hardy joined Mercury Interactive Software as vice president of global accounts and was instrumental in driving growth resulting in the acquisition by HP. Hardy also served as vice president of worldwide sales at Xsigo, which was subsequently acquired by Oracle. Most recently, Hardy served as vice president of global accounts and service providers at Blue Coat Systems

“As enterprises and their customers continue growing around mobile and cloud computing, they will need the kind of high performance arrays WHIPTAIL provides to ensure their processes are executing quickly and efficiently,” said Hardy. “WHIPTAIL is changing the way these organizations build their infrastructure for optimum performance and scalability. I look forward to joining a team that is at the forefront of the flash storage space, continually innovating as the industry evolves.”

The post WHIPTAIL Appoints Former EMC Executive Erik Hardy as Executive Vice President of Worldwide Sales appeared first on WHIPTAIL.

• Database
• Batch
• Email
• Online Transaction Processing (OLTP)
• Analytics/Intelligence
• High Performance Computing
• Video

The post Webinar: 20 Minute WHIPTAIL Overview appeared first on WHIPTAIL.

The post TechValidate Research Results appeared first on WHIPTAIL.

The post Webinar: Bottom Line Benefits–How Flash is Saving the IT Budget appeared first on WHIPTAIL.

• PostgreSQL and PG_Bulkload
• Oracle 11g and Datapump
• Microsoft SQL Server and Bulk Insert

To find the answers, we ran bulk load tests using each database in our environment, built upon an INVICTA all-flash storage array.

The Tests
WHIPTAIL Labs began by loading 1,000 rows of census-like data into a single table in a database on a single LUN. We then scaled the dataset by a factor of 10 for each subsequent load until we reached 100,000,000 rows loaded. We completed these tests for Oracle, PostgreSQL and Microsoft SQL Server; you can see our results here.

Results

Observations

• As evident in our results, your performance will depend upon the capabilities of your application
• Fine tuning applications for performance storage will help to take advantage of IO and throughput
• The more aggressive the application is at loading data the more capable it is of handling large data levels

Download the notes as a PDF.

The post Bulk Load Scaling Datasets: PostgreSQL, Oracle and MS SQL appeared first on WHIPTAIL.

I’d like to talk BIG DATA, albeit “from a certain point of view” (yes, that’s a Alec Guinness via Obi-Wan Kenobi, Return of the Jedi reference). At best, BIG DATA is a nebulous and overused term. To draw from well-known statistics that I’ve just made up (though believable): for every organization successfully leveraging some conventional BIG DATA workload–Hadoop or some other platform or parallel file system, a non-relational DB, a J2EE n-tier architecture, etc.—there’s another three organizations that are still actively exploring the potential benefits of such solutions and, beyond that, at least 10 more organizations that collectively groan with each mere mention of the term.

While I’m as indulgent as the next geek in discussions concerning the future of technology, I still tend to think like a former integrator—more focused on providing end-users best in breed experience w/ the current state of solutions and, consequently, would tend to count myself among the latter group of people who are tired of the industry’s emphasis on solutions that seek to cater to very specific needs or to a fringe set of organizations sometime between now and the distant (yet, somehow, always foreseeable) future.

This position is certainly not a knock against shops sitting atop or behind a massive parallel file systems or n-tier structures, as there are giant service providers and familiar e-marketplaces among these organizations that fuel and guide the tech industry and the retail space (and, in some cases, both). That said, I’ve recently been musing that, beyond the lighthouse accounts and/or industry titans that are leveraging hyperscale solutions today, there’s got to be something in this BIG DATA conversation for all of us.

I once had a chemistry professor who threatened to lead each exam with the following confused jabberwocky: “Tell everything you know about the universe, true or false, and list two examples.” With this preamble and a slightly less cryptic scope in mind, I’ll cite three.

What, Me QUERY?
For nearly as long as I’ve been in my current role, I’ve been working with a customer that I’d describe loosely as a fleet management and logistics outfit. The organization monitors and/or manages the activity of a significant number of fleet vehicles on behalf of its customer base. This operation requires the ingest of a non-trivial amount of data from numerous agents attached to field vehicles, collecting specific metrics at regular polling intervals and feeding centralized data hubs. Generally, this process should not be unfamiliar to anyone within the logistics or the intelligence communities. Due largely to the I/O constraints, the organization established SLAs that provided customers the output of real-time or near-time queries of this data during production hours, but longer-term queries that relied on larger sets of historical data were intended to be scheduled during off-hours, ensuring that heavier query execution would not hinder production data collection.

Ultimately, conversations and solutions that changed the way this organization viewed I/O-heavy workloads yielded a corresponding change in the company’s thoughts towards line of business applications and its service offering to its growing customer base. Initially, the provider considered a freer, more on-demand approach to historical reporting. During more recent discussions, the organization expressed interest in expanded data collection metrics and reduced polling intervals, contemplating the perceived value and competitive advantage of these enhancements in the eyes of existing and potential customers.

Keep Calm and BATCH On
I’ve been engaged with another logistics service provider with similar processes, though geared towards rail management rather than fleet, for over a year. As above, this organization encountered obstacles on the back-end of a large data set. Query execution presented bottlenecks that slowed the availability of current (and, inherently, more valuable) data, which consequently hindered internal communication and exchanges with customers. In my experience, this is a story common among business intelligence solutions: a batch operation that commences off-hours, intended to complete during the course of night, does not finish until sometime during the middle of the next business day.

The above is a potential Hydra-headed phenomenon: internal teams enter the workday without current and meaningful data and rely instead on stale information, compute and I/O heavy workloads associated with the batch operation are now running during production hours and potentially affecting productivity, and other critical off-hour routines (AV, backup, etc.) or maintenance may be deprioritized to meet the needs of the business. In this instance, the organization succeeded in reducing these batch operations to a two-hour window by eliminating I/O constraints.

All the WRITE Moves
A third organization with a borderline-oppressive amount of data faced challenges on the other side of the data flow—the front-end ingest rate. A marketing and lead generation service provider, managing messaging campaigns for a growing customer base, was tasked with meeting the demands of a transaction-heavy workload. A single campaign may generate upwards of a million messages per hour, and, correspondingly, each active customer represents significant increase in the overall transaction rate, potentially resulting in significant I/O contention. Despite legacy storage solutions that were backed by flash and a non-trivial number of spindles, this I/O contention constrained resources available to the storage processors, yielding dangerously high latencies approaching three-figures (in milliseconds).

By eliminating potential I/O bottlenecks and reducing these latencies to sub-millisecond values, this service provider was able to cultivate existing customers and pursue new business without fear of disproportionate growth to back-end assets and resources to meet exponentially increasing I/O demands. Customer-facing SLAs are no longer at risk, internal help desk calls are near-nil, and general-purpose infrastructure workloads are seeing enhanced performance on a now unburdened legacy storage solution—the veritable converse of the many-headed, serpentine problem previously described.

Enter the EVERYMAN (Cue the Aaron Copland)
There’s commonality among these three examples that create an interesting context in which to discuss the immediate future of data needs and potential data growth:

1. Outside the earshot of insiders to each respective industry, none of these shops is a household name. The largest organization has employees numbering in the low four-figures, whereas the other two are currently sitting at around or just below three-figures. Regardless of name-recognition and headcount, all of these shops struggled with and successfully resolved issues that stem from data at high-scale.
2. Each is relying on in-house applications to meet the needs of business-facing demands. None are currently leveraging parallel file systems or non-relational DBs like MongoDB, and instead they’re relying on more traditional DBs like MS SQL and Oracle. More succinctly, none of these organizations needed to overhaul line of business processes or data structures to overcome immediate challenges.

In much the way former Alaska governor Sarah Palin sought to appeal to “hockey moms” nation-wide, I feel that it’s time for an earnest appeal to BIG DATA for the Everyman, an exploration not merely of the potentially transformative power of near-term and/or emerging solutions (which I do not deny, deployed appropriately) but rather to the pressing problems of rampant data growth and the ever-present need to efficiently and intelligently glean meaningful output from these datasets.

The post “Everyman’s” BIG DATA, or How I Learned to Stop Worrying and Mine the Data appeared first on WHIPTAIL.

Recent initiatives aim to develop systems that allow government agencies to garner insight and understanding from the massive amounts of information being collected at the Federal level. Your goal is to unleash the potential of your data faster and find actionable results quicker but budget restraints are affecting everyone. At the crossroads of lofty Big Data goals and limited funding is the storage decision.

Chris Nichols covers:

• Architecting Storage for Big Data
• Accelerating Mission Critical Workloads
• Analyzing More Data, Faster
• Getting Actionable Information Quicker

The post Big Data in the Federal Sector appeared first on WHIPTAIL.

The post WHIPTAIL Offers Clues to Users Playing the SSD Box Riddle Game appeared first on WHIPTAIL.

The post Agile Company – Federal Government appeared first on WHIPTAIL.

Recent initiatives aim to develop systems that allow government agencies to garner insight and understanding from the massive amounts of information being collected at the Federal level. Your goal is to unleash the potential of your data faster and find actionable results quicker, BUT budget restraints are affecting everyone. At the crossroads of lofty Big Data goals and limited funding is the storage decision.

Join us for a webinar, hosted by Chris Nichols, that will cover:

• Architecting Storage for Big Data
• Accelerating Mission Critical Workloads
• Analyzing More Data, Faster
• Getting Actionable Information Quicker

The post Webinar: Big Data in the Federal Sector appeared first on WHIPTAIL.

• Overcome I/O bottlenecks using solid-state flash devices
• Combine flash technology with existing disk drives
• Prevent storage from slowing down or taking down your applications

This is a great opportunity to build your professional network, share best practices, and receive firsthand knowledge of successful strategies and solutions from your peers. Register here.

The post Ignite Performance by Combining Flash with Storage Virtualization appeared first on WHIPTAIL.

The post WHIPTAIL Technologies Scales Flash Array Down to SMB appeared first on WHIPTAIL.

The Goal
WHIPTAIL LABS, tasked with rethinking the deployment of real-world applications to better exploit the performance characteristics of WHIPTAIL flash arrays, asked ourselves these questions:

• How does PG_Bulkload compare to Copy?
• How do we improve PostgreSQL throughput while using very little memory?
• How fast can it load 15 Billion rows of data?

To find the answers, we ran hundreds of bulk load tests, keeping lab notes of our process and the results.

Five Easy Steps
1. Build Your Environment

• Two Brocade 300s
  8GB/24 Port
• Fibre Channel Connections
• Xeon E7 Intel Based Server
  24 Logical CPUs
  112GB Ram
  Qlogic 2Port 8GB HBA

2. Attach Your Storage

• WHIPTAIL INVICTA All-Flash Array
  Dual Routers
  Four 6 TB Nodes
• Create a LUN on each Node

3. Configure Your OS – PostgreSQL | Linux 6.3

• MPIO
  path_grouping_policy: multibus
  /*Active/Active*/
• Linux Boot
  2.6.39-300.28.1.el6uek.x86_64
  numa=off
  elevator=noop
• Disks
  nr_requests: 4096
  read_ahead_kb: 4096

4. Configure Your Database – PostgreSQL

• Default Settings
  32 MB Database Cache
• 1 Tablespace per LUN
  5 Tables per Tablespace
  System Activity on Tablespace 1

5. Load Using PG_Bulkload

• Set Parallel Option in the Control File

Test Environment

4 Node INVICTA Configuration

• 425,000 IOPs
• 4.7 GB/s Sustained Throughput
• 200 Microsecond Latency

The Tests
WHIPTAIL Labs ran the PostgreSQL copy command for a single 100GB file (1 Billion rows of data with 11 unique attributes) reading from one INVICTA node to another multiple times. We then ran the same tests using PG_Bulkload to see the difference in the amount time it took. We proceeded to expand the PostgreSQL instance to four tables on a single node, loading them from the original source file concurrently. Finally we scaled our database to 15 tables spread across 3 nodes for a total of 15 billion rows being processed at once from a single source file. You can see our results on the next page.

Results

The Take Away
The first thing we noticed was the difference in time it took pg_bulkload to process the same amount of data when compared to using PostgreSQL copy. We were encouraged to find out how quickly we could load multiple tables concurrently using pg_bulkload.

Unlike our Oracle Lab Note, in which we tested the effect of increasing parallel processes with Datapump, Pg_bulkload was not able to drive as much data through the system. We instead increased the number of tables to drive throughput and designed these tests to demonstrate the ability for pg_bulkload to write to multiple tables per LUN.

In our tests, it was clear that load time did not increase proportionate to the volume of data; we were able to load 15 billion rows of data in under twice the time it took to load 1 billion using pg_bulkload and in less time than it took copy to load 1 billion rows. Taking full advantage of the INVICTA scale out architecture by distributing multiple tables across the array, we were able to fully demonstrate the scale out ability of pg_bulkload.

Observations

• Pg_Bulkload logs every operation; accounting for about 70% of the sustained write activity in our results.
• If we were able to run tests without log operations, we propose that this would cut the total load time in half.
• We did run tests with larger cache sizes, but didn’t notice a substantial change in throughput, so all results here are shown using 32MB Ram
• Because read performance didn’t increase as we increased the number of tables, we were limited with how aggressively we could write to our tables
• We noticed that the elapsed time documented in pg_bulkload did not always match up with the CPU time tracked
• All testing was done using opensource software and file formats (.csv), not proprietary versions.

The post How PostgreSQL Scales Out to Load 15 Billion Records at Once appeared first on WHIPTAIL.

The post Big Data, Storage and the Need for Speed appeared first on WHIPTAIL.

There are an increasing number of high-performance storage providers, particularly in the all-flash space, that are beginning to offer higher-density or clustered solutions. Nearly all of the these solutions are subject to limitations stemming from similar resource constraints. These all-flash clustering solutions, heralded as scale-out platforms, rely on redundant controllers or network gateways to ensure high-availability for a set number of drive shelves, and the performance output of these solutions is typically a function of the throughput capabilities of these controllers. WHIPTAIL’s INFINITY architecture is the first solution that provides end-user organizations the ability to expand storage clusters beyond the typical controller-pair model to ensure that a cluster’s performance continues to scale in a linear fashion even beyond the limits prevalent in the marketplace.
-Paul Costello, Director of Engineering

Clearly, WHIPTAIL feels we’re offering an incredible solution with INFINITY. Why? INVICTA INFINITY is the first all-flash storage array built on the idea of massive I/O parallelism throughout the entire design of the array. It takes the already proven methodology of INVICTA‘s scale-out platform and allows it to grow as far and as fast as your needs in your data center do. Simply put, it allows you to once and for all answer the questions, “What if I no longer ever had to worry about storage as a performance constraint? How much differently could we do business if we could accelerate our systems to operate at real-time, all the time?”

Time and time again, we speak with customers telling a story of massively over-provisioning storage resources to address a performance need. But, because today’s storage systems all have bottlenecks at scale, this becomes a shell game of how to use those limited I/O resources to cause the least pain. With the INFINITY architecture, these issues disappear. It is the first and only all-flash array available today to be designed for the capabilities of flash storage at scale, and to do so in an entirely modular approach. Storage bottlenecks will no longer be an issue in your data center, for both performance AND capacity needs–a concept that, up until now, was always an either/or conversation with respect to storage.

Time and time again we’ve proved our solutions operate at the speed of business. INVICTA INFINITY has the potential to deliver impact to your business to instantaneously change HOW you do business.
-Chris Nichols, Director of Field Engineering of the Americas

In data centres today, storage systems struggle to cope with capacity additions and linear performance growth, with INVICTA INFINITY this goes away. WHIPTAIL are delivering solutions that put the user back in charge of when and where to deploy the technology. When a customer plans for the year ahead they don’t have to make massive buying decisions at the start of the year and can deploy changes throughout the years as project timelines become more concrete and performance variables are better understood. With the ability to plug in extra performance and capacity when required the customer regains control of the out of control storage spends that they have been used too in the past with upfront and then knee jerk buying decisions common. With the INFINITY architecture they know they can keep adding capacity and performance only when required and not have to worry about plateauing performance and bottlenecks.
-Darren Williams, EMEA/APAC Director of Operations and Sales Engineering

The post Engineer Sound Off: INVICTA INFINITY appeared first on WHIPTAIL.

Whippany, N.J. – June 4, 2013 – WHIPTAIL, the data storage innovator powering faster and more energy-efficient computing for today’s global businesses via flash storage, today announced the WT-1100, the newest addition to the company’s portfolio of flash-based silicon arrays. Designed for small and medium businesses (SMBs) or companies with branch offices, this addition gives customers the highest levels of choice and flexibility to scale up or scale out their application-intensive environments.

WT-1100 Enables Faster Turnkey Solutions for SMBs and Branch Offices
This low-profile, high-performance 1U solid state storage array features an installation wizard that speeds deployments, making it ideal for turnkey implementations focused on virtual desktop infrastructure (VDI), email and databases. The WT-1100 runs WHIPTAIL’s RACERUNNER operating system, which optimizes the write performance of NAND flash. The product will be sold through authorized WHIPTAIL partners alongside other WHIPTAIL products like ACCELA and INVICTA. The array will begin shipping in Q3 2013.

“Our partners’ success depends on their ability to deliver high performance solutions that consistently perform well over time,” said Dan Crain, CEO of WHIPTAIL. “By creating the WT-1100 we are helping our partners solve big problems for small businesses with the same quality and performance once reserved for large enterprises.”

Expansion Gives Flexibility and Choice in Deployment Options
As an entry-level complement to WHIPTAIL’s impressive arsenal of flash-based storage solutions, the WT-1100 features up to 4TB of SanDisk SSD capacity, 100,000 IOPS, and <0.1ms latency in a 1U chassis. The configuration wizard simplifies installation while RACERUNNER provides optimized performance and endurance of flash.

“Organizations of all sizes need affordable access to flash storage technology to accelerate server and storage performance,” said Steve Fingerhut, Vice President, Marketing, SanDisk Enterprise Division. “The SanDisk Enterprise SSDs deployed in the WT-1100 offer high predictable performance, trusted reliability and superior value. SanDisk enables the WT-1100 to pack enterprise level technology into a cost-effective price point, which SMBs and remote offices can fully utilize to drive their businesses.”

ACCELA is a 2U energy efficient NAND flash silicon storage array providing up to 12TB capacity while delivering 250,000 IOPS and 2GB/s throughput, which dramatically improves application response time. ACCELA supports many workloads including databases, virtual desktop infrastructures, web, email, analytics and online transactions.

INVICTA INFINITY Will Be Generally Available in July
Announced back in November 2012, INFINITY is a major expansion to WHIPTAIL’s flagship INVICTA Modular Storage Array that will scale the six-node 72TB INVICTA array up to 30 nodes and up to 360TB of all flash storage. A fully populated 30 node, 360TB INFINITY will exceed 4 million IOPS and 40GB/second throughput in real world use. INFINITY will be available to enterprise customers in July.

With INFINITY, enterprise customers can scale performance and capacity as needed, by simply adding another 12TB module. This allows for the consolidation of high performance, high demanding applications without compromise onto a single platform.

The post WHIPTAIL Delivers New Level of Performance for Partners Servicing SMB Markets appeared first on WHIPTAIL.

The scenario: You have a billion rows of data in a file that you need to extract, transform and load to an Oracle database and only one LUN available. Not only do you need to process that database quickly, but you know that the next challenge you face is to process three equally large files at the same time without affecting the total time it takes to load.

This recorded webinar walks through the results of our tests. Matt Goldensohn, Director of WHIPTAIL Labs, and Max Riggsbee, VP of Product Management, will discuss the lessons learned in the process, including:

• Data loader efficiency
• Critical Linux IO performance settings
• Why real data is preferable over synthetic IO generators for testing
• Trouble shooting bottlenecks
• How to scale data without taking a performance hit

The post ETL Nirvana appeared first on WHIPTAIL.

To learn more, read Data Center Modernization Starts with Cloud, SDN and Virtualization.

The post Poll: Which Technology is Your Company Using to Handle Computing Demands? appeared first on WHIPTAIL.

The scenario: You have a billion rows of data in a file that you need to extract, transform and load to an Oracle database and only one LUN available. Not only do you need to process that database quickly, but you know that the next challenge you face is to process three equally large files at the same time without affecting the total time it takes to load.

Join us for this webinar that walks attendees through the results of our tests and allows for questions following the presentation. Matt Goldensohn, Director of WHIPTAIL Labs, and Max Riggsbee, VP of Product Management, will discuss the lessons learned in the process, including:

• Data loader efficiency
• Critical Linux IO performance settings
• Why real data is preferable over synthetic IO generators for testing
• Trouble shooting bottlenecks
• How to scale data without taking a performance hit

The post Webinar: ETL Nirvana appeared first on WHIPTAIL.

The post How to Get Your VDI Storage Right appeared first on WHIPTAIL.

More than three years after shipping our first product, we felt it was time to take a closer look at the WHIPTAIL effect. We partnered with TechValidate to survey hundreds of customers across two dozen vertical markets, each running a variety of workloads on our technology. What we found was beyond our expectations. Yes, customers are able to run workloads faster and, as a result, they are finding a competitive advantage in their market. Some have described business process improvements as high as 300%. Others are spending less time managing technology and more time deploying business solutions. Many are eliminating performance tuning and shortening QA cycles. And almost all are experiencing savings. Here are a few highlights from the results:

In the coming weeks we will be releasing more of the TechValidate results. We thank our customers for selecting us to help them drive business forward. Our TechValidate survey highlights what happens when you solve the write problem.

The post TechValidate Research Results appeared first on WHIPTAIL.

The post Whiteboard: Factors that Affect Flash appeared first on WHIPTAIL.

Video Transcript
Hi, my name is Matt Goldensohn and I’m the Director of WHIPTAIL Labs. We’re here today to discuss things that affect flash and, specifically, things that you would be concerned with when you’re trying to figure out what kind of storage array you’re looking for or how well it does. We’re going to talk a little bit about things that you’re going to run into regardless of the vendor or the manufacturer or how the array is put together. These are the kinds of questions that you should ask about what the array is doing, where it does it and how well it does it.

What

To get into this quickly, key things . . . Write amplification is the process where you exceed your erase block. Meaning, if your erase block is 2MB in size then you’re going to write 2.2MB or something larger than 2MB in erase block pages. You will amplify your writes by actually requiring two pages or doubling your writes to 4MB. Pretty simple. Write amplification is, I’m larger than an erase block or page.

Garbage collection is the process by which we free up erased blocks. In other words, when we’ve written something and moved that data somewhere else or invalidated the old page we have to go back and erase the old or invalid data. That process is referred to as garbage collection.

Over provisioning is a pool of pages or capacity that every solid-state storage device will contain to do garbage collection. In other words, it’s not capacity necessarily that’s addressable by you, but it’s used during the garbage collection process. Every SSD, or solid-state device, will have some space set aside for over provisioning, which is used by the garbage collection process.

BTL and FTL is block translation layer or flash translation layer. It is the technique by which we address data and pages in the storage array or at the solid-state device layer. Most BTLs employ a log structure file system or LFS. We’ll talk a little bit about that.

Wear leveling and mean time to failure (MTTF) is a function of all the above.

And dedupe and compression is also something that dramatically impacts performance. By dramatically, it takes longer if you’re going to check for whether or not something already exists, therefore it’s a duplicate, or compress it. So when I go to write, typical write performance is affected by lengthening the amount of time that it takes to write because I have to check for duplicates or I have to compress the data. On the write performance, I have to do those checks on the way in. On the read, if I’ve compressed it, I have to decompress in order to present the data back to you. Those [dedupe and compression] are separate topics and we could probably spend a good amount of time just on that alone. But I want you to be aware of it.

Where

The next thing that we really need to focus on with respect to this is where does this stuff take place? Typically, all of these functions take place on a storage processor. At least with most of the arrays. I’ve got an onboard controller here, but I’ll get to that in a minute. Your storage processor is going to be the layer, or processing layer, which deals with all of these steps. Nodes, or your storage node, may have some compute power on them and may have some intelligence and may be able to do this as well. If it’s not handled at the node layer, every commercially available solid-state storage device has an onboard controller. That onboard controller can do all of these with the exception of dedupe.

The where is really important with respect to scale. If we’re doing garbage collection, for example, on the solid-state device itself then we are, one, limited by the processing capability of that onboard controller. Therefore, things will take longer because that will not scale well; you will always have to check for free pages on that individual device. In general, that’s something you never want to have happen. As a matter of fact, most of this stuff you really don’t want to take place on the onboard controller. If you are doing things well, they’re all taking place at the storage processor layer.

Let’s talk about that for a little bit because with respect to arrays, there’s two types of arrays that you can really put together. One array is a scale-up array, which means that as I need capacity, I can add more drives or add more nodes. I’m not necessarily increasing the number of storage processors, but I am increasing the capacity that I have to store data and potentially the amount of data that I’m going to store in that array. The nodes themselves, in that particular case, are not actually handling any of the compute power; that compute is all staying at the processor or the controller up here at the top of the array.

If I’m distributing my processing power down to the node, I can move some of these tasks, the execution of those tasks, off of my storage processor and out to the actual node themselves. The advantage of that is when I increase capacity I’m also increasing my ability to do things, such as dedupe, such as garbage collection, and deal with things at the node layer.

Last but not least is this onboard controller, which is your fallback regardless of where you’re doing your actual compute. Again, worst case is to have to rely on this. But when evaluating these things, find out where each of these processes takes place because that will truly impact the performance of your overall array.

When you’re looking arrays, you got to try to figure out, okay, they all do these things but what does that mean to me, how do I measure it, how do I know if your array does this better than another array, what’s the difference.

How

There’s really two metrics that I have at the end of the day that are impacted by all of these–time and capacity.

Overprovisioning, the over pool of available pages that I have for garbage collection is a capacity function. Everything else is time. Meaning that if I’m doing a good job at garbage collection, I won’t degrade my time. So I can measure that. If my garbage collection algorithm is executing as it should, then I will see consistent write and read performance from my array. If my garbage collection process is not working well, then I will see degradation in my write and read performance. Meaning, it will take me longer to read and write, and it can be pretty significant and measurable in terms of the amount of times.

It’s not just write. What’s going to happen is the onboard controller, or somewhere in here, you’re going to have to say, woah hold on. I’ve got to empty some of these pages. I’ve got to clear up some more space so that I can do garbage collection. And until I do that, I cannot return your IO. I can’t give you your data or I can’t write your data. So time function.

Your BTL and FTL are also time function measurable. If I’ve got a good [BTL or FTL], or if I’ve done it well, I’ll see really good read and write performance. And by that I mean they’ll be identical, or close to identical, within time. That’s key. Typically, we measure write and read performance in number of IOPs and we associate a block with it. You could have a 4K block, an 8K block and I will do 100,000 IOPs, I’ll do 500,000 IOPs or some measurable amount of reads and writes within a period of time. Then, what I want to make sure is that if I’m doing, say, 250,000 read IOPs, they’re random, they’re 4K, if my BTL–my block translation layer and my log-structured file system are working well–my write performance should be similar to that, if not identical.

How well I’m doing is how close are those numbers. What’s the ratio? If I can read 250,000 IOPS at 4K blocks then I should be able to write the same. That is key in flash because flash typically does not do that. So flash typically requires that you erase before you write new data. It adds another step. If you’ve dealt with that in your BTL or in your log-structured file system where I’m always appending new pages, then you’ll see equivalent time performances. The metric here is time. So how well am I doing depends on what kind of time responses I’m getting.

Again, when you’re looking at this stuff, ask these questions: How many IOPS do you do at 4K? Random? What’s your read-write percentage? If I can do 500,000 IOPS at 4K write, what’s my read? I’m looking for balance. That tells me how good of a job I’m doing.

And last but not least is compression and dedupe. In addition to all of these things which I have to do, what’s the impact of your de-duplication or compression algorithms? How much more time do I need for that write IOP to take place?

Conclusion

When we’re looking at evaluating these technologies, you have to figure out what to look for.

• Where is that being done?
• Is it being done at the storage processor layer? Is it done at the onboard controller layer?
• How do I measure the impact of how well these tasks are being executed?

It should be pretty straightforward. You should be able to develop a set of metrics that I can walk back and say, hey, I know over time, when I wrote so much data, that I didn’t see any degradation; I haven’t noticed that my writes and reads have degraded in time. So there’s no indication that I’m doing garbage collection. Or I can see, for example, that my writes and reads are equivalent. I’ve got some sense of equal in terms of I can do so many writes and I can do so many reads. If I can do far less reads, that’s a function of an imbalance in the operating system or the way that I’m implementing these challenges.

The post Whiteboard: Factors that Affect Flash appeared first on WHIPTAIL.

Understanding risk is difficult; it requires huge pools of data to be analysed at speed. As these data sets increase, and underwriting becomes more personalized, this challenge only grows. Consequently, the London insurance market is driving some of the most demanding analytics workloads as, evermore, it is the key to profitability, customer satisfaction and cost control.

Analytics, databases and VDI are mission critical applications that require affordable, high-performance storage. Time and again insurance customers are outgrowing their storage performance needs within their investment cycle; WHIPTAIL was founded to deliver unparalleled storage performance whilst dramatically lowering the cost of delivering these applications. Whether it is accelerating front line workloads or delivering VDI to employees, discover how companies are improving their services whilst lowering their costs – doing more with less – with WHIPTAIL.

Following this webinar a subset of attendees will be able to put a WHIPTAIL array through its paces in their own test environment. We are confident you won’t be disappointed.

The post Webinar: Time to Information is Critical for Insurance, Analyse More – Faster appeared first on WHIPTAIL.

The post Poll: How Are You Using the Cloud for Personal Storage? appeared first on WHIPTAIL.

The post Data Center Utopia: Accelerate Applications While Simplifying Your Data Center appeared first on WHIPTAIL.

In the current economic climate, successful construction businesses have increasingly invested in software solutions to automate manual processes (e.g., estimating, job cost accounting, project management) and increase revenue without driving up overhead costs and putting the bottom line at risk. Decision makers are evaluating existing technology to determine whether the company is leveraging all features and functionality to ensure optimal productivity and profitability.

Advancements in communications have made information about the construction market and changing regulations instantly available in a form that can be effectively assimilated and used for decision making and now enable the field crew to access updated project data via a mobile device. Contractors can quickly and accurately answer customer questions while at the jobsite, saving significant time. Anyone on the team can easily produce, organize, route and manage project documents and can create alerts to identify overdue items and meet deadlines.

As a rapidly maturing technology, many in the construction industry have turned to virtualization to keep a lid on spiraling IT costs without losing control over their data. More than ever, the enterprise data center is essential to provide a strong and secure base for information and knowledge storage and distribution.

Running a Virtual Desktop Infrastructure (VDI) and database applications on WHIPTAIL’s high-performance, energy-efficient silicon storage arrays allow servers to process more data in less time. WHIPTAIL’s Flash-based silicon storage solutions offer unsurpassed levels of performance and ultra-low latency that ensures rapid response times for processing and real-time analysis of industry trends, while supporting hundreds to thousands of users simultaneously. WHIPTAIL’s performance metrics are perfectly suited for VDI and far exceed traditional HDD (100 times the IOPS). WHIPTAIL’s products prevent current and future performance issues caused by bottlenecks and resource contention by removing VDI’s high demand IO profile from the general storage infrastructure.

Operational efficiency is the name of the construction game. WHIPTAIL products dramatically accelerate the delivery of data, provide 90% savings for on-going energy costs and offer a 70% reduction in data center footprint as opposed to traditional magnetic spinning disk storage. We’ve created products that are easily attainable for an IT infrastructure in relation to the amount of performance that results from the investment.

By optimizing existing technology solutions with WHIPTAIL, construction executives can run their businesses more effectively, streamline processes, reduce expenses and anticipate profit in a difficult economy.

The post Construction appeared first on WHIPTAIL.

To further that effort, we’ve launched two communities: one in Google and the other in LinkedIn. Both are called the Application Design and Deployment in Flash Environment groups. Now these groups are not about us. These groups are about how do we go about rethinking the deployment of applications for flash.

Inside of these groups we’re looking to foster a whole set of conversations, really look at the way customers in the past have organized their applications to, say, avoid writing because it was painful, and now what we can do to engage it because it’s becomes easy. Really begin to think how we can consolidate databases or, even more importantly in some cases, consolidate a process inside of a database, i.e., let’s do a batch while doing a backup for example. So I encourage you and welcome you to join us in these two communities. We look forward to your comments and your conversations. Thank you.

The post Announcing the Application Design and Deployment in Flash Environments Groups appeared first on WHIPTAIL.

The Goal
WHIPTAIL Labs, tasked with rethinking the deployment of real-world applications to better exploit the performance characteristics of WHIPTAIL flash arrays, asked themselves three questions:
• How much data throughput can a single instance of Oracle support?
• How fast can it load multiple tables from a single file?
• How fast can it load multiple tables from multiple files?

5 Easy Steps
1. Build Your Environment

• Two Brocade 300s
  8GB/24 Port
• Fibre Channel Connections
• Xeon E7 Intel Based Server
  32 Logical CPUs
  128GB Ram
  Qlogic 2Port 8GB HBA

2. Attach Your Storage

• WHIPTAIL INVICTA All-Flash Array
  Dual Routers
  Four 6 TB Nodes
• Create a Lun on each Node

3. Configure Your OS – Oracle | Linux 6.3

• MPIO
  path_grouping_policy: multibus
  /*Active/Active*/
• Linux Boot
  2.6.39-300.28.1.el6uek.x86_64
  numa=off
  elevator=noop
• Disks
  nr_requests: 4096
  read_ahead_kb: 4096

4. Configure Your Database – Oracle 11g

• Default Settings Except for
  Asynch IO
• 1 Tablespace per Lun
  Oracle System Activity on Tablespace 1

5. Load Using Data Pump for Oracle

The Test
WHIPTAIL Labs ran data pump for a single 100GB file (1 billion rows of data with 11 unique attributes) reading from one INVICTA node to another ten times. They then expanded the Oracle instance to two table loads on two different nodes, reading from the same file concurrently, again running it multiple times. Next they added a third node and third table load for a total of 3 billion rows being processed concurrently from a single source file. But they weren’t done yet, the lab re-ran their tests with multiple source files, again building from 100GB to 300GB total. You can see the results on the next page.

Test Environment

Results

*A peak throughput with Oracle was observed at 5.6GB and documented with 5GB/s Write and 600MB/s Read concurrently.

The Take Away
You will notice that the time to load did not increase proportionate to the increase in the volume of data. In fact, relative to typical load times for files 100GB or larger, the amount of time it took to load additional data was minor.

This is due in part to WHIPTAIL’s scale-out architecture. By distributing the flash management, power, and RAID resources on each node, INVICTA is able to give dedicated service to each tablespace within the Oracle instance. That means there is zero conflict for IOPs or throughput, and Oracle is able to scale to meet the growing demands of business.

For the database administrator the impact is significant. INVICTA creates the ability to consolidate large databases under a single Oracle instance without concern for degrading performance.

Observations and Tips

• In the first three tests, we observed that as the number of writers requesting data from the same file increased, the read throughput decreased. Because the writers were asking for the same data, requests could be serviced directly from cache and therefor the read throughput from the file itself was lower.
• In the fourth run, loading from two files to two tablespaces, there was a marked improvement in overall run time and throughput, as the system had sufficient resources to service concurrent read & write requests.
• By the final run, it became evident that our environment was constrained by the operating system on the server. Linux simply did not have enough resources to read and write three large files simultaneously.
• To remove that constraint, you could opt to architect your solution with additional resources.
• The question to ask yourself is; When was the last time your storage was able to push the boundaries of your processors?

The post How to Bulk Load 3 Billion Rows of Data in About 5 Minutes appeared first on WHIPTAIL.

Today we’re very happy to launch the first of a series of notes that will appear on our website. In this particular note we’re going to demonstrate to you the ability to load billions of records in just a matter of minutes. It’s not that these are really small records–we’re talking about 100 gigabytes of data for about a billion records. Those records represent something that resembles census information, and so this will give you an opportunity to understand how you could organize a database and get extraordinary throughput. Not all databases are made the same and the notes will reflect that as well. What you’ll begin to find in today’s notes, for example, is about two to three pages outlining the steps that we took, the systems that we used, how we organized the array, which is very simple as a matter of fact, and how we organized the database around the array, and then the tool that we used, which you would be able to use in order to execute the load.

Now here’s the best part. We’re also going to make available to you the output of our tests. You’ll be able to see the log files that were generated, IO stat information, any screen captures that we’ve done and we’ll make that available. And, we’re doing this because we want to begin to solicit a whole new way of thinking about how applications can be organized specifically for flash. We also want your comments. We want to hear what you think, and certainly we want to know what kinds of notes you’d like to see. In the coming weeks you will see us do work on Oracle, Microsoft, Postgres, MongoDb. We’re looking into Hadoop. We’re looking at doing work with NoSQL. We’re looking at doing work with all kinds of other systems, and so we’re very excited about this initiative. We’re very excited to share it with you, and now on to the first note.

The post Beyond the Benchmarks with WHIPTAIL Lab Notes appeared first on WHIPTAIL.

The post WHIPTAIL CEO: Big Blue’s Flashy BEELLION Doesn’t Faze Us appeared first on WHIPTAIL.

Register here.

The post Webinar: Data Center Utopia – Accelerate Applications While Simplifying Your Data Center appeared first on WHIPTAIL.

Throughout the webinar we polled attendees on their thoughts about Gary’s TCO model, if they are currently using flash and predictions for the HDD performance market. Here are the results:

After reviewing the results, Gary said, “The results of our TCO survey indicate to me that storage buyers still do not have a handle on how to justify the extra upfront cost of an all-flash array vs. traditional HDD for a performance environment. Purchase price is only one dimension of storage cost. Keeping HDD available and running over a period of 3 to 5 years is much more costly than a comparably powered flash array and offsets any upfront savings.”

Let us know your thoughts in the comments and by voting in the below polls:

The post Webinar Recap: Calculating the True TCO for Performance Storage appeared first on WHIPTAIL.

Let us know your thoughts by voting in the below polls:

The post Calculating the True TCO for Performance Storage appeared first on WHIPTAIL.

Join WHIPTAIL’s CMO Max Riggsbee and Parity Research Founder Gary MacFadden as they discuss Parity’s ground-breaking research on the true total cost of ownership (TCO) for performance-centric storage arrays. This 30 minute webinar, plus Q&A, will cover key findings from MacFadden’s white paper, Taking a Closer Look at Data Storage TCO.

Key findings to be covered include:

• Why HDD arrays cost almost twice as much to maintain over 5 years than a comparable SSD array
• Why HDD arrays take up 8 times more physical space than a Flash array
• How HDD arrays consume almost 10 times the power of Flash arrays
• Why administrative costs are almost 500% higher with HDD
• Why HDD deployment costs are more than 5 times the cost of deploying SSD Flash.
• How One 6-terabyte SSD Flash Array can replace 480 short-stroked HDDs with 50x reduction in latency
• Why capacity and IOPS are inadequate tools for measuring performance-centric data storage TCO

The post Webinar: Calculating the True TCO for Performance Storage appeared first on WHIPTAIL.

What if your Exchange, SQL and Oracle Databases, VDI, Server Virtualization, and Mission Critical Revenue Generating Application workloads could operate faster to handle today’s demands and tomorrow’s expectations?

Let Cornerstone and WHIPTAIL show you how to reap the benefits of today’s solid-state storage. Join us for a luncheon workshop and learn about solid-state storage arrays that offer the performance and scalability to help drive your business to its full potential.

We will cover:
·Increasing demand for data velocity
·Righting the IO imbalance
·Accelerating Workloads
·Customer Stories
·Q&A Session

Register here.

The post Cornerstone Luncheon Workshop appeared first on WHIPTAIL.

The post Flash Storage a Win for Delaware, Ohio appeared first on WHIPTAIL.

Symmetry Affects Performance
Writes are important, more important than many realize. One of the most common practices in storage design is to say, “OK, let me look at my application and its IO profile.” This is a critical step, but its results are often misinterpreted. The goal is to understand how that application works, how much time does it write, how much time does it read and so on for the entire lifecycle of the data. Some applications spend more time doing one activity versus the other and we often make assumptions regarding the importance of write IOPS if our write activity is less than 50%.

Let’s say I’ve done my IO evaluation and I’ve discovered my application writes 30% of the time and reads 70%. I may think write symmetry with respect to performance does not matter, but that is not actually the case. I have to consider the criticality of that particular write operation.

As a hypothetical IO analysis let’s consider a banking application that produces account activity statements at the end of each month for all customers, which is required by bank regulations. The application is going to go through everybody’s accounts, total them up, summarize them and store that summarized data in the form of the month-end statement. For most of the month the database of record is writing far less of the time than it is reading.

Let’s dig in to the data collection used by the application.  If I have 50 million accounts with an average of 25-50 transactions per month, I get 1,250 to 2,500 million transactions created each month.  Now consider that if you keep 12 months worth of transactions per year in your current customer data collection the average of at least the last fiscal years’ transactions could total between 30 to 60 trillion records.  The program to calculate month-end statements will be doing a lot of reading. It will focus on the current month’s data, which is 1/12th of 30-60 trillion records. But, more importantly, it will be doing extremely critical writes in an extremely short period of time.

Symmetry Implies Quality
Read/write symmetry is important to look at when considering flash storage beyond just making sure applications can perform write operations quickly. As previously discussed, flash array vendors use a block translation layer (BTL) to manipulate and improve the write performance of flash. This BTL does much more than increase the speed at which flash writes. But if an array has read numbers that far exceed the write numbers and one sees a dramatic degradation in write performance with that array, then you can safely infer that storage solution software layer is carrying the write asymmetry back out to the operating system of the host application and the application itself.  This ratio of write to read performance should be looked at closely when evaluating an SSD array regardless of the IO Profile of an individual application.

Write performance isn’t only important to applications, it’s essential to the entire business. Applications cannot read anything that hasn’t first been written and they can only go as fast as their slowest write, which means the entire business may be at the mercy of its storage as it waits on writes.

The post Why Read/Write Symmetry Matters for Your Business appeared first on WHIPTAIL.

Following this webinar a subset of attendees will be able to put a WHIPTAIL array through its paces in their own test environment. We are confident you won’t be disappointed. In addition, one attendee will be able to take advantage of a free application health check, carried out by an independent 3rd party, to understand if storage performance is holding your business back.

The post Webinar: Financial Services – Take Control of Legacy Storage Performance and Costs appeared first on WHIPTAIL.

WHIPTAIL: What compelled you to want to research and compare the TCO of HDD and Flash?
Gary MacFadden: I like a challenge. I have worked with and seen a number of storage total cost of ownership (TCO) models over the years. But none of the models I’ve looked at made head-to-head comparisons between all-hard disk drive (HDD) and all-Flash arrays when implementing and sustaining performance-centric, enterprise-grade data storage assets. Most storage TCO models focus on capacity or IOPS, both of which are insufficient measurements for arriving at a true total cost of ownership for performance-centric data storage. To project a realistic TCO one would have to include environmental, labor, maintenance and storage management software costs.

Other performance-centric storage TCO models I’ve seen include server core reduction and lower costs for non-storage-related software licenses while leaving out labor or storage administration costs. To me, fewer servers or database licenses is one of the many return on investment (ROI) benefits of Flash, which could also include faster response time, improved employee or user satisfaction, shrinking backup windows or the ability to support new business initiatives. The ROI for implementing Flash arrays is enormous, but it doesn’t tell the CFO or the IT department how much to budget for storage only. A pure storage TCO model does that.

WT: How are you qualified to conduct this type of research?
GM: I’ve been part of the IT research and advisory industry for more than twenty years. I was introduced to TCO modeling about 15 years ago and over the years have worked with major corporations to help them negotiate multi-million dollar storage and server purchases. Interestingly, before I joined META Group in 1990, which was bought by Gartner in 2005, I sold route accounting and inventory management software for handheld computers that were among the first applications for solid state disk (SSD). Almost 25 years later, it’s amazing how far the technology has advanced.

WT: What did you expect to find? 
GM: When discussing the scope of this project with Max (Maxwell Riggsbee, WHIPTAIL CMO), he convinced me that a storage TCO model focused on performance would also demonstrate how Flash was more cost effective than traditional HDD. Frankly, I figured that even if the costs were equal that would be a compelling story – especially because of all the potential downstream benefits, including faster processing, lower energy costs and a smaller data center footprint.

WT: How did your expectations compare to the results? 
GM: I was very surprised. You might say shocked. When I created the formulas that sit behind the TCO model, I gave every advantage possible to HDD so as not to be biased toward Flash. For instance, the model assumes each individual Flash drive and HDD consume the same amount of power even though HDDs are known to consume more. The model also assumes that drive failure rates, ongoing maintenance, training and provisioning costs are equal. I included the highest performing HDD, measured in IOPS, with the deepest discounted prices I could find.

Over five years, a mid-side WHIPTAIL performance array is almost half the price of a comparably powered HDD array. The big savings for Flash customers are in power and cooling, data center rack space, and deployment and ongoing administrative costs. Simply put, all-Flash arrays are easier to deploy than Tiered HDD solutions or hybrid solutions that use a combination of Flash and HDD. Reduce complexity and you lower cost.

WT: Why do you think people are hesitant to believe Flash storage is more cost effective for performance purposes? 
GM: I think that hesitancy is grounded in the fact that many organizations still rely on traditional TCO models that use storage capacity and IOPS as the primary measurement when calculating costs for performance storage. It could also be that some organizations don’t consolidate or aggregate their costs the way our TCO model does. For instance, if the initial purchase price for data storage hits the IT budget but the ongoing facilities costs do not, then that organization will have a skewed view of overall storage costs. It’s as if you bought a car to commute 100 miles a week to your job and only factored in the purchase price of the automobile. You also need to account for gas, maintenance, license, insurance and perhaps even property taxes, tolls and parking costs.

WT: How do you think your findings affect the industry? 
GM: The word “disruptive” comes to mind. There is already enough anecdotal evidence that Flash is much less expensive – not to mention much faster – when performance is the primary goal. As a result, the market for enterprise-grade, performance HDD is collapsing. I believe Parity’s TCO work on performance-centric data storage is ground-breaking because it explicitly spells out all of the true storage TCO components and the model is simple enough that any application owner, purchasing or finance person can understand and easily adopt it for their own use.

The post Gary MacFadden Discusses His Data Storage TCO Research appeared first on WHIPTAIL.

The post GUI Shorts: Creating a LUN appeared first on WHIPTAIL.

The post WHIPTAIL Product Deep Dive appeared first on WHIPTAIL.

The post GUI Shorts: Creating Users and Roles appeared first on WHIPTAIL.

Rethink your storage strategy and prevent VDI nightmares from affecting your users’ experience while you save yourself from constantly fine-tuning your infrastructure. What if you could run as many anti-virus scans as needed throughout the day? What if your storage array could handle a boot storm of 600 desktops while concurrently managing constant database queries and an anti-virus scan, all without lag time or the dreaded spinning pinwheel? This webinar covers VDI nightmares, how to prevent them and what these changes mean for storage administrators in their day-to-day functions.

If you would like to learn more about the calculator referenced in this webinar, please contact us to set up a one-on-one conversation to review our VDI Calculator.

The post How to Prevent VDI Nightmares appeared first on WHIPTAIL.

Rethink your storage strategy and prevent VDI nightmares from affecting your users’ experience while you save yourself from constantly fine-tuning your infrastructure. What if you could run as many anti-virus scans as needed throughout the day? What if your storage array could handle a boot storm of 600 desktops while concurrently managing constant database queries and an anti-virus scan, all without lag time or the dreaded spinning pinwheel?

Join us for a webinar covering VDI nightmares and how to prevent them. We will also address what these changes mean for the storage administrator in their day-to-day function.

Register here.

The post Webinar: How to Prevent VDI Nightmares appeared first on WHIPTAIL.

The post GUI Shorts: Main Dashboard appeared first on WHIPTAIL.

Darren Williams, WHIPTAIL SE Director of EMEA and APAC, will be speaking at the show on Tuesday and Wednesday.

Storage and High Performance – Making it Possible
23rd April 2013 at 11:30AM
Darren will explain why traditional storage is not good for performance, how NAND flash should be managed and how WHIPTAIL has eliminated the weakness from NAND in terms of endurance, write speed and affordability with its operating system.

Batch Processing – Why it’s Traditionally Done at Night and Why it Takes So Long
24th April 2013 at 3PM
Darren looks at how users supported by traditional storage have to run batch jobs overnight in order to have them finish before the start of the business day. This process means information can be out of date by the time users can access it. Darren will let attendees know the ramifications of this practice and that there is another way.

The post The Big Data Show 2013 appeared first on WHIPTAIL.

While working as the Director of Infrastructure at a financial institution we had an end-of-day report that was taking longer and longer to run as the business was picking up.  The amount of data read by the report was growing and the calculations needed to generate interest rate sensitivity curves were completing into the next business day.  At some point we either had to shorten the report production step or choose not to take on any new business. We took the process, which had been running on a single production cluster, and distributed it across multiple server clusters and partitioned the data to mitigate the read and write latencies that were extending the report process. It was a less than ideal workaround needed because of the disk storage’s slow performance, but it allowed us to continue growing.

Disk vs. Time
There are time penalties associated with both hard disk and flash storage. When writing, HDDs spend some of this time just making sure the drive’s head is in the right position on the track and the platter has spun to the correct location. These are referred to as rotational and seek latency and it potentially adds six milliseconds of time.

As a workaround, database administrators have employed many techniques such as over-provisioning, the act of buying more storage than is needed; short stroking, only storing data on a small portion of the disk; and caching, writing temporarily to a non-persistent memory before backing up to persistent storage. While these techniques do bring minimal time savings, they are all very costly to a business and waste investments.

Of these techniques the most commonly used is to execute as much IO in non-persistent memory as possible.  This type of memory typically can perform IOs in a millionth of the time needed to write to hard drives.  While the speed will allow your applications to feel as if they’re running faster, at some point you have to make the data persist.

Flash vs. Time
This leads us to NAND flash, which is organized into blocks that are then divided into pages—eliminating rotational and seek latency. Flash media executes reads and writes a thousand times faster than rotational storage. However, this speed boost is not enough and unless the IO to flash media is manipulated, flash writes data much slower than it reads.

For example, if I were to buy a thumb drive from a computer supply store and plug it into my laptop, the writes to flash used in that drive are managed solely by the on-board controller and my write activity is going to be slower than my read activity. Why is this? An extra step is required when writing to flash; it must erase the block it is going to write to before it can write the data.  If you are writing less than a block, it reads the existing data, erases the data and then writes the merged data back.

Enterprise flash storage vendors have developed techniques to mitigate the problem of slow writes by adding a block translation layer (BTL). Each vendor’s BTL operates differently and finds varying results based on what techniques are used, but one of the goals is to equalize read and write performance.

As a part of BTL, a storage array will perform a function called “garbage collection.” This process uses a “lazy erase” to erase all of the flash memory pages that the array has read data from before writing it somewhere else so that the erase did not have to be executed at the time of the write. The process eliminates the extra step and increases the array’s write performance.

To fit appropriately in the data center, flash must be manipulated. And to find optimal storage performance, businesses must compare the read and write ratio offered by flash arrays. In my next blog I’ll explain why this read/write symmetry is so important.

The post It’s a Matter of Time: How Disk and Flash Battle Time in Search of Performance appeared first on WHIPTAIL.

The post Performance Centric Data Storage TCO Study Demonstrates Cost Advantages of Flash Arrays appeared first on WHIPTAIL.

The post Webinar: WHIPTAIL Product Family Deep Dive appeared first on WHIPTAIL.

• One to Watch Award Product
• One to Watch Award Company
• Flash/SSD Product of the Year

Traditionally, the readers of Storage Magazine reward outstanding products, services and people in the IT industry. The 2012 Awards experienced a new record of 8.500 readers who voted for their favorite products, people and companies and that number is expected to grow. Vote for WHIPTAIL at the Storries’s website.

The post The Storries Nomination appeared first on WHIPTAIL.

UPDATE: Want to learn more about MacFadden’s findings and ask him questions about his research? Attend the Calculating the True TCO for Performance Storage webinar.

 To embed this infographic on your site:
<a href=”http://whiptail.com/blog/infographic-what-is-the-tco-for-performance-storage/”><img src=”http://whiptail.com/wp-content/uploads/2013/04/What-is-the-TCO-for-Performance-Storage.jpg” alt=”What is the TCO for Performance Storage – Infographic” title=”What is the TCO for Performance Storage – Infographic” width=”825″ height=”4344″ /></a><br/><a href=”http://whiptail.com”>WHIPTAIL – Performance Storage</a>

The post Infographic: What is the TCO for Performance Storage appeared first on WHIPTAIL.

We encourage you to make your voices heard and vote for WHIPTAIL! We are in the SSD Flash Product of the Year category (found under Technology Awards) here.

The post Data Centre Solutions Awards Nomination appeared first on WHIPTAIL.

The post When Performance Counts, Think Flash: Taking a Closer Look at Data Storage TCO appeared first on WHIPTAIL.

Hard Disk Drive Storage & Read/Write Symmetry
With spinning magnetic media, data is written to or read from an addressable location on that drive, which is either a sector or a portion of a sector residing in a track that sits on a disk. In order to perform either action, the heads are moved to the proper location on the disk and the data is written or read respectively. Because either action takes the same amount of time, hard disk drives always offer symmetric reads and writes, but they also experience rotational and seek latency.

Memory Storage & Read/Write Symmetry
Memory can be broken down into two classes: persistent and non-persistent. With non-persistent memory, a block of data is written at a specified address to a value and data is read from that same block of bytes. Similar to HDDs, non-persistent memory has symmetric read/write behavior. If non-persistent memory storage loses power the bytes of data are lost and cannot be recovered or accessed again.

Persistent memory, the predominant version of which is NAND flash, is asymmetrical. It reads data in blocks. Its write activity is a system of adding redundant (parity) data to a block such that it can be recovered. This occurs in four steps. It reads the existing data in the block, merges it with the new data, erases the original data and then writes the new data to the block. These are specific and unique operations not present in HDDs or non-persistent memory. Because of this, the write operation takes much longer to perform than the read operation and read/write symmetry can only be achieved when the flash write is manipulated.

Achieving read/write symmetry should be a priority for your business. However, this does not mean HDD storage or non-persistent memory is always the right option. The pros and cons of each solution must be weighed while also evaluating how a storage vendor has manipulated its flash arrays and what read/write numbers it is then able to attain. I’ll be discussing this topic more in upcoming posts.

The post Read/ Write Symmetry: What is it and What Storage Options Have It appeared first on WHIPTAIL.

The post Six Hot Storage Startups appeared first on WHIPTAIL.

In order to mine and analyze data by performing multi-variant hypothesis testing on complex systems or by performing predictive modeling for market perceptions, the IT challenge is to find the best solution for reaching an ever expanding user base with a corresponding demand for timely, high-performance access to targeted data.

Imagine the positive effect if systems could process analytic intelligence and generate reports for extracting and integrating data 10 times faster. WHIPTAIL’s family of products provides a solution for companies struggling with the ramifications of moving data fast enough in virtualized environments, Big Data applications and continued pressure to make rapid decisions.

inVentiv Health implemented a faster virtual environment with WHIPTAIL. “After deploying WHIPTAIL’s silicon storage solution we have found our virtual environment is running 4X faster and the help desk calls have stopped completely,” stated Tony Marando, Manager of Technical. “Those reports that took hours now only take minutes… The 10X increase in improved performance offers us a distinct competitive advantage. WHIPTAIL’s underlying Flash-based architecture shows tremendous performance. We were able to stop over-provisioning in spindles (while wasting capacity) to try to optimize the underlying storage.”

The post Analytics & Intelligence appeared first on WHIPTAIL.

• Increasing demand for data velocity
• Righting the IO imbalance
• Accelerating Workloads
• The WHIPTAIL Product Family
• Customer Stories
• Q&A Session

Tweet about and during the webinar using #PropelBusiness

The post Webinar: How WHIPTAIL Accelerates Application Workloads and Propels Business appeared first on WHIPTAIL.

James Candelaria, CTO
SaaS, PaaS and IaaS are really interesting and extremely powerful ideas that, for very different reasons, are beginning to take hold in both small-medium and large enterprises.

The small and medium businesses out there tend to have fewer resources to dedicate specifically to specializations, which allows them to offload some of their requirements to the Cloud or to experts more capable of handling those solutions. For the enterprise it’s more about scale—providing burst capability for workloads that may not be able to work efficiently inside of their own data centers.

Flash is an important driver in use cases both small and large.  Small businesses leverage larger providers such as Amazon, who has introduced new high IO capabilities that utilize unmanaged flash to present a best effort SLA. The larger organizations using SaaS, PaaS or IaaS will also need to leverage flash, but will want to take a more strategic approach.  This will require a more robust managed flash infrastructure that can dynamically grow or shrink the performance envelope as necessary to meet the business goals/SLAs while not falling victim to the “noisy neighbor” issue.  For the providers, a truly scalable, managed flash platform is required to make this objective a reality.

Maxwell Riggsbee, VP of Product Management and CMO
The appeal of SaaS, PaaS and IaaS is speed of deployment. Companies move to IaaS because the infrastructure is ready to use.  One can immediately take an application, put it onto the platform and get going. The same is true for SaaS. One does not have to build a CRM system; they can simply start using a preexisting one such as SalesForce.com.

This movement is fueled by companies that face slow application deployments due to hard disk drive technology, which requires a lot of configuring and maintenance. Flash starts to change that. I believe a number of applications that have moved to IaaS will move back to in-house IT as flash enables a simpler and faster way to deploy applications and eliminates the need to configure hard disk drive technology.

Accelerating applications is the most important thing for IaaS and SaaS customers. Expect to see as-a-service providers adopting flash to fuel performance.

Ryan Snell, VP and General Manager of the Americas
Most businesses of all sizes already take advantage of SaaS today (Salesforce.com, Concur, Office365, etc.).  This trend should continue, as not only is the risk lower than ever, but there is a dramatic increase in competition driving down costs.

Regarding PaaS and IaaS, it’s hard to see short-term adoption grow past large enterprise development departments and large enterprises with strategic platform projects that see unanticipated surges in infrastructure requirements.  Ultimately, security will remain the number one obstacle while complexity and cost will follow thereafter.

The post Executive Sound Off: As-a-Service appeared first on WHIPTAIL.

The post WHIPTAIL Breaks Vendor Lock-In On Storage Replication appeared first on WHIPTAIL.

WHIPTAIL’s INVICTA is a next-generation NAND flash storage array that provides highly scalable IOPS, total system throughput, and ultra-low latency at groundbreaking price for performance, while ACCELA offers an unsurpassed high-performance solid-state data storage solution.

Installed worldwide in numerous mission-critical database, virtualisation, and online environments, WHIPTAIL’s solid-state storage dramatically reduces delays related to hard disk contention and access times, allowing servers to process more data in dramatically less time – months to days, hours to minutes.

First Distribution has strong representation by its vendors in the storage arena, having well-set and established vendor relationships with IBM, EMC, Oracle, Huawei, and Iomega. The latest agreement with WHIPTAIL adds further strength and depth to the company’s already impressive storage portfolio.

Having taken note of the rapidly-emerging industry trend towards big data, and the accompanying storage needs, First Distribution has shored up its storage division accordingly. Solid-state drives are a highly promising emerging technology, and integrate well with First Distribution’s reputation for distributing innovative and industry-leading products and solutions.

Ludwig Dedekind, General Manager of First Distribution, commented that “flash storage is everywhere these days, and while it’s common in mobile devices, it’s now also gaining traction within the traditional storage sphere of high-performance desktop computers, servers, and enterprise-scale storage systems. I think we’re going to start seeing it becoming increasingly attractive for businesses to switch to flash memory, as it does not have the mechanical limitations of hard drives – as well as being faster and less noisy and consuming less power. Flash storage technology is also continuing to improve in maturity, cost, and reliability. One of the features which I really admire about WHIPTAIL is its use of innovative data acceleration software to bring enterprise-level I/O speeds to users without any of the hardware degradation that these speeds usually entail.”

Ludwig added that “WHIPTAIL’s flash-based NAND silicon storage is exactly the kind of leading technology which First Distribution prides itself on distributing. We’re very pleased to have signed the agreement with them, and look forward to working together with them in the future.”

Brian Feller, VP and General Manager of WHIPTAIL EMEA, sees uptake of Flash-based storage in enterprise environments on the rise: “IT managers are realising that traditional spinning disks are no longer able to deliver the quality of end-user experience expected when it’s ‘business as usual’ at stake. Big data applications and continuing pressure to make rapid decisions mean companies are struggling to move data fast enough. There is strong demand for technology to increase performance in these areas.

“Organisations are increasingly turning to flash storage to provide the resilience required in performance-intensive, mission-critical environments. Flash has historically been perceived as prohibitively expensive, with short lifecycles and lower write capacities than traditonal storage – but we are shattering those myths and providing First Distribution’s customers with the ability to dramatically cut their storage total cost of ownership while moving data faster than ever,” he concluded.

The post WHIPTAIL Appoints First Distribution as Southern African Distributor appeared first on WHIPTAIL.

March 26, 2013 – Hedensted Kommune has selected WHIPTAIL’s performance storage technology to power 800 virtual desktops, including mobile devices, across the local authority’s network over the next year. Hedensted Kommune has migrated 400 desktops to WHIPTAIL’s storage array, and has saved approximately 1.5 million kr (approx. $225k) in immediate storage costs.

With an objective to increase staff productivity as well as provide an improved service to citizens, Hedensted Kommune’s front line workforce is increasingly mobile. Employees require access to both its SAP business management solution as well as the business critical data held within it, at any time, including from tablets and smartphones. Hedensted Kommune researched the benefits of a virtual desktop infrastructure (VDI) to enable mobile workers to run the data-hungry application.

“I was very focused on the performance problems caused by running VDI on hard disk drives,” commented Kim Bengtsson, Networks and Server Administrator at Hedensted Kommune. “The field employees cannot afford to wait on data reaching their devices; we need to maintain a quality service and high levels of productivity. Running VDI on hard drives or an SAN would make the project unfeasible on a cost and performance level, but WHIPTAIL’s storage array would deliver the performance that we required.”

Bengtsson was first introduced to WHIPTAIL at Citrix Synergy in Barcelona, through distributor Commaxx. He said: “WHIPTAIL was the missing link I was looking for – I was concerned about the initial cost of silicon storage technology but they showed us how to run VDI the way it should be, and at a very cost effective price point.”

Hedensted Kommune has migrated from a hard-drive based storage area network (SAN) to Citrix XenDesktop, supported by a 6TB WHIPTAIL storage array installed as the authority’s local server. According to Bengtsson: “Pre-rollout stress tests on the WHIPTAIL array have reached an impressive 250k input/outputs a second. During migration I have not yet been anywhere near pushing the box to its limits. This is no doubt the best piece of hardware I have ever had out of the box; anybody can configure it within 25 minutes – and that includes a coffee break.”

The local authority has already saved 1.5m kr by choosing to deploy its 800 virtual desktops on WHIPTAIL’s silicon storage array, rather than a hard-drive based SAN.

These incredible savings are sustainable, as Brian Feller, VP and General Manager EMEA at WHIPTAIL added: “With public sector budgets continually under the microscope, organisations are under pressure to find efficiencies where possible. Hedensted Kommune can expect to save money every time an employee access SAP from their mobile device, as data is sent and received in lightning speed. VDI demands high input/outputs for it to work effectively, plus Hedensted Kommune can expect to easily scale up as it grows the workforce.”

Bengtsson said: “Our next project will be to run VDI with personal Vdisk for our graphics-demanding users – as we are confident that WHIPTAIL’s technology will be able to deliver the required IOPS and low latency for this scenario as well.”

The post Hedensted Kommune Launches New Virtualisation Platform on WHIPTAIL appeared first on WHIPTAIL.

“I love WHIPTAIL, it’s so freaking fast and it never goes down.”

These were the first words that greeted me when I sat down with Fred Schimscheimer, designer of the RAWC (Desktop Reference Architecture Workload Code) and senior technical marketing manager at VMware’s Palo Alto HQ. Fred’s had a WHIPTAIL for nearly a year now and is one of our most stalwart supporters on their product marketing team. While my visit helped me recap with Fred, its main purpose was to explore an upcoming joint project between WHIPTAIL and VMware.

A large part of my role in WHIPTAIL’s Business Development team is working with Independent Software Vendors (ISV’s) to understand their applications and how we can closely work together to create a solution that provides great benefits to our mutual customers. This could be as simple as making their back end database faster or as complex as co-operatively coding with them to offload some of the application’s storage requirements direct to the WHIPTAIL box. VMware’s VAAI is a perfect example of this.

For ISVs, the mean time to market with the latest version of their application is key, but always has to be tempered by releasing a robust and bug-free product.   This is achieved by multiple parallel tests, Quality Assurance (QA) test runs, simulated user tests and internal code releases…. And all of this is before we even get to the Alpha and Beta test phases! The process is resource intensive and time-consuming as well as frenetic as the company races to release the latest features into a crowded marketplace. Being the first to market with a long sought-after feature gives a massive competitive advantage for the organisation that manages it.

Fred created the RAWC tool in order to generate a realistic, adjustable user workload simulation for VDI projects, the results of which can be analysed to identify the appropriate environment for a VDI project. The RAWC workload runs on a Windows 7 or Windows XP operating system and opens applications such as Word, Outlook and Adobe Reader and even initiates a Virus Scan at random to mimic the actions of a desktop user.

Fred ran me through his latest project and whilst I can’t divulge the exact nature of the tests, I can tell you that it is a competitive comparison between two VDI-based products and it is sufficient to note that 1/100^th of a second is an important degree of measurement and accuracy is key, especially as the test scales outwards to several thousand users. When I probed a little further about why Fred chose WHIPTAIL for his current project he replied, “Transparency in your testing is vital and you want it so that your storage isn’t leaving you wondering whether it’s impacting your results or not. WHIPTAIL was the only storage I trusted to give me that.”

VMware isn’t the only ISV whose testing has been impacted by its storage. Both large multi-discipline organisations and dedicated ISVs have found dramatic improvements when moving their storage to flash. I am always interested in hearing from ISVs regarding what they’re working on and performance issues they may be facing. Find me on Twitter, LinkedIn or share your thoughts in the comments below.

The post Why Storage Matters for ISV Testing appeared first on WHIPTAIL.

What would happen if a critical batch job, especially one that interfaces with other critical functions, does not complete within its window? Is there a danger of breaching a Service Level Agreement that can result in significant financial loss or a compromised reputation in the industry?

High-volume batch processing puts heavy demands on systems. WHIPTAIL products add write and read performance and scalability to enable batch processing without pausing workflow.

One WHIPTAIL logistics customer in the freight transportation industry was experiencing major operational inefficiencies–although they depended on round-the-clock collection of customer messages, they needed to shut down their messaging system for six hours each night to run batches. When the batches completed, the start of regular workday activities was significantly delayed until the backlog of overnight messages could be received, filtered and addressed. With WHIPTAIL, batch processing was reduced to two hours and no longer requires shutting down the customer messaging system.

The post Batch Processing appeared first on WHIPTAIL.

The post From Big Data to Fast Data in a Flash appeared first on WHIPTAIL.

For most companies, as more workloads demand more performance they respond by purchasing more mechanical hard disk storage. The result? Marginally better workload acceleration that drives IT effort whilst consuming more energy and more floor space. Mechanical hard disk drives constrain workloads and business agility.

Matt Goldensohn, Director of Business Solutions, presents another option. Rethinking the storage decision allows accelerating workloads to drive business effectiveness rather than IT effort. He reviews real world examples of multi-workload testing with solid-state arrays and the resulting metrics.

The post Achieving Extraordinary Workload Performance appeared first on WHIPTAIL.

The post Storage and High Performance–Making it Possible appeared first on WHIPTAIL.

The post Flash: Enabling the ‘Mission-Critical Business’ appeared first on WHIPTAIL.

For most companies, as more workloads demand more performance they respond by purchasing more mechanical hard disk storage. The result? Marginally better workload acceleration that drives IT effort whilst consuming more energy and more floor space. Mechanical hard disk drives constrain workloads and business agility.

Join Matt Goldensohn, Director of Business Solutions for WHIPTAIL, as he presents another option. Rethinking the storage decision allows accelerating workloads to drive business effectiveness rather than IT effort. He will discuss real world examples of multi-workload testing with solid-state arrays and the resulting metrics, as well as answer your questions about multi-tenancy on flash storage.

Register here.

The post Webinar: Achieving Extraordinary Workload Performance with Solid-State Storage appeared first on WHIPTAIL.

The post In a Spin appeared first on WHIPTAIL.

The post Global Customer Support appeared first on WHIPTAIL.

Join us for a webinar covering best practices for deploying VDI on flash storage. Discover how WHIPTAIL’s ACCELA and INVICTA storage arrays provide the optimal platform for running virtual desktops and allow businesses to deploy thousands of desktops within a single 2U shelf of storage.

Register here.

The post Webinar: Best Practices for VDI Storage in the Legal Sector appeared first on WHIPTAIL.

What if…

• Storage performance limitations were the furthest thing from your mind?
• VDI environment could scale to thousands of users without any performance concerns?
• Separate reporting and processing silos could be eliminated because the impact on production was no longer an issue?
• Virtual tier 1 workloads could out accelerate traditional physical workloads?
• Most importantly; what if this was at a cost already in line with today’s expectations?

Simply put, it’s an enormous shift in thinking and it’s a critical shift in storage design for virtual infrastructures. Attend this event and learn how WHIPTAIL helps its customers achieve incredible results by employing our flash storage solutions. Register here.

Attendees will be entered to win a $100 gift certificate.

The post Lunch with Broadleaf and WHIPTAIL appeared first on WHIPTAIL.

Tweet about and during the webinar using #WHIPTAILin20.

The post Webinar: 20 Minute WHIPTAIL Overview appeared first on WHIPTAIL.

The post Multi Workload Demo appeared first on WHIPTAIL.

Tweet about and during the summit using #BigDataFast.

Register here.

The post Webinar: From Big Data to Fast Data in a Flash appeared first on WHIPTAIL.

The post Infographic: A Lot Can Happen in a Second appeared first on WHIPTAIL.

Whippany, N.J. – March 12, 2013 – WHIPTAIL, the data storage industry innovator powering faster and more energy-efficient computing for today’s global businesses via flash storage, announced today that Cristóbal Conde has been named chairman of its board of directors.

“Having Cris on our board has been a boon for our young company, and as we mature, having him as chairman will continue to pay dividends for WHIPTAIL,” said CEO Dan Crain. “His experience and advice has been invaluable to our continued growth.”

Conde is a senior advisor with TPG Capital and Providence Equity Partners, serves on the boards of Digital Reasoning Systems and MassMutual Financial Group and is a trustee for the College of Mount Saint Vincent. He comes from a background in data centers and is considered a pioneer in guiding financial and energy companies to solutions for complex business problems from his days as president and CEO of SunGard.

“From power to performance to footprint, the modern data center is feeling more pressure than ever,” said Conde. “That pressure is accelerating the usage of memory-based storage and WHIPTAIL’s all-flash arrays are designed from the ground up to eliminate all three pain points. It’s an elegant solution to a growing problem.”

In addition to Conde, WHIPTAIL’s board of directors includes Richard Fade of Ignition Partners, Avi Falks and Raymond Wong from Spring Mountain Capital and Will Porteous from RRE Ventures.

The post Conde Named WHIPTAIL Chairman of the Board appeared first on WHIPTAIL.

Watch this archived webinar to examine the trends affecting the need for application speed and how each company’s storage decision will either constrain or unleash its most critical workloads.

Hosted By:
Max Riggsbee
VP of Product Management and CMO, WHIPTAIL
@mkriggsbee

Tim Stammers
Senior Analyst, 451 Research
@carryonpompei

The post Storage Trends appeared first on WHIPTAIL.

Join us for a webinar that examines the trends affecting the need for application speed and how each company’s storage decision will either constrain or unleash its most critical workloads. Register here.

Hosted By:
Max Riggsbee
VP of Product Management and CMO, WHIPTAIL

Tim Stammers
Senior Analyst, 451 Research

The post Webinar: Storage Trends 2013 appeared first on WHIPTAIL.

There always seems to be confusion when talking about IOPs and virtual desktops (VDI). I often have people tell me their users do 40-50 IOPs. My next question is “what is the read/write ratio?”, which is often followed by a deer-in-head-lights look. What does it mean to say a user does 40 IOPs? Is that total IOPs or IOPs to storage and does it include user data IOPs? And how do you quickly determine how many IOPs you will need? Physical desktops are easy, all reads and writes go to the physical device. It is not that simple for VDI configurations, even dedicated persistent VMs, due to raid configurations of data center storage and the capability to separate read and write IOPs to different locations.

Some definitions to help clarify:

Raid configurations will not affect the number of read IOPs per user, but can greatly affect the number of write IOPs per user. A Raid 1 Mirrored configuration requires two physical writes for every user write IOP and is most commonly used by HDD vendors to obtain optimal performance at a capacity cost. Raid 5 requires 4 IOPs (read old parity, read old data, write new data, write new parity) for each user write IOP, gaining capacity at the cost of performance. To achieve acceptable performance using Raid 5 requires using SSD arrays.

For VDI configurations not doing fully dedicated persistent VMs, conventional best practices have read and write IOPs separated to allow for optimal performance. Read performance and the user experience will be affected if reads and writes go to the same storage.

Both Citrix XenDesktop and VMware View offer best practices on separating reads and writes. Citrix XenDesktop utilizes PVS (Provisioning Server) to manage the master image on a separate server. Reads go to the provisioning server and, if configured to Citrix best practices, writes will go to the storage utilizing write cache files for each VM. VMware View allows for placement of the replicas on a different location than the storage containing the associated linked clones. The Replica is placed on high performing/low capacity storage like SSDs to provide optimal read performance. The writes are to the linked clones on higher capacity storage. With read and write IOPs separated the read/write ratio becomes important.

If a user is doing 40 IOPs, but the reads come from one location and the writes go to a different storage location, we need to break down the IOPs. As an example, assume a user is 40/60 read/write (r/w) ratio. This means 40% (16) of the IOPs are read, and 60% (24) are writes. So, using the Citrix model we can know the reads are coming from the provisioning server and the only writes are going to our storage. But what about user data? A best practice for VDI has the user data storage on a different storage location like a share and not within the VM. As a rule of thumb, user data IOPs are a very small percentage of IO and are often in peaks such as the opening of a file. Until the file is saved or closed nothing is written back to the user data storage, however temporary data is written every few seconds to the write cache/differential files, which is on the storage we are concerned about. To account for user data IO, a rule of thumb is to specify 1 IOP each for read/write of user data, so in our 40 IOPs user example accounting for user data the numbers become 15 read IOPs and 23 write IOPs allowing 1 read and 1 write IOP for user data.

In our example, the 15 read IOPs would come from the provisioning server and the 23 write IOPs would go to our storage. However, testing has shown that in a Citrix XenDesktop configuration using a PVS provisioning server and write cache files, the r/w ratio to the storage is actually 10/90. There are some reads that come from the temp files on the storage. A rule of thumb here is to allow 1 read for every 10 writes, always rounding up.

To determine the IOPs to our storage:
Storage IOPs = Write IOPs + Read IOPs
Storage Write IOPs = ((Total IOPs * write %) – 1) * Raid Penalty
Storage Read IOPs = (Total IOPs * write %) / 10 rounded up

For example, our 40 IOP user has a 40/60 r/w ratio, and we will assume a Raid 1 mirror. The calculation would be:
Storage Write IOPs = ((40 *.6) – 1) * 2 = (24 – 1) * 2 = 23 * 2 = 46
Storage Read IOPs = (40 * .6) / 10 = 24 / 10 = 3 rounded up
Total Storage IOPs = 46 + 3 = 49

To support 1000 users would require 49000 IOPs from the storage if using Citrix XenDesktop and provisioning server handling the read IOPs.

If doing VMware View and following the best practice of placing the Replica(s) on a different location then the calculations are the same as with Citrix XenDesktop.

However, if the replicas are on the same storage as the linked clones, then View will perform very similar to a fully dedicated persistent VDI session. For these scenarios the Storage IOPs would be determined by:
Storage Write IOPs = ((Total IOPs * write %) – 1) * Raid penalty
Storage Read IOPs = (Total IOPs * read %) – 1

As an example, a user doing 40 IOPs with a 40/60 r/w ratio moved to a fully dedicated persistent VDI session using Raid 1:
Total IOPs = 40
Storage Write IOPs = ((40 *.6)-1) * 2 = (24-1) * 2 = 23 * 2 = 46
Storage Read IOPs = (40 * .4) – 1 = 16 – 1 = 15
Total Storage IOPs = 46 + 15 = 61

These formulas are rules of thumb and are designed to help estimate IOP requirements. To fully understand IOP requirements it is recommended an assessment of your users be done.

The post IOPs and Virtual Desktops appeared first on WHIPTAIL.

The post About WHIPTAIL appeared first on WHIPTAIL.

The post How Storage Technology Works appeared first on WHIPTAIL.

Find WHIPTAIL at booth #1103 between 8 AM and 6 PM.

The post VMware Partner Exchange 2013 appeared first on WHIPTAIL.

But according to Storage-Switzerland.com, WHIPTAIL’s modular all-flash arrays bring a cost effective way to deal with rapidly increasing IOPS needs and introduce a high-performance, scalable solution into the data center. Read Colm Keegan’s Briefing Note to learn more.

The post Colm Keegan’s Briefing Note appeared first on WHIPTAIL.

The post Liquidware Labs Launches New Program for Technology Alliance Partners appeared first on WHIPTAIL.

Rapidly Expanding Customer Base Leads to Outsourced Manufacturing with Leading Provider of Purpose-Built Application Platforms, Appliances

WHIPPANY, N.J. – Feb. 21, 2012 – WHIPTAIL the data storage industry innovator powering faster and more energy-efficient computing for today’s global businesses via its innovative flash storage arrays, has partnered with global manufacturer NEI to meet demand for WHIPTAIL’s family of products.

NEI provides application platforms, deployment solutions and lifecycle support services for software technology developers and OEMs worldwide. It is known for its expertise in storage, security and communications. WHIPTAIL chose NEI as its manufacturing partner because the company has extensive experience in solution design, manufacture, global logistics and lifecycle management. By leveraging NEI’s qualified procedures and process, WHIPTAIL has been able to speed time to market for its storage arrays while minimizing the effort and resources that WHIPTAIL would otherwise expend, enabling it to concentrate on its core business.

“We’ve been shipping product for a little more than two years now, and have broken the 200 customer threshold,” said WHIPTAIL CEO Dan Crain. “When we were a small company, we could handle our own manufacturing, but that’s not the case anymore. NEI has a commitment to quality workmanship and very efficient business operations. They take pride in delivering and supporting a high-quality, reliable product, and that’s what our customers expect.”

For NEI, adding WHIPTAIL is another validation of its LEAD (Lifecycle Engine for Application Deployment) System, which includes all steps from design, integration, management, logistics and more. NEI’s LEAD model is perfectly suited to help software developers like WHIPTAIL deliver and support powerful, affordable appliance solutions to meet their schedules and competitive performance criteria.

“We have a unique skill set in this industry that enables young companies to economically and efficiently bring their products to market and compete with established players on a level playing field,” said Jeff Hudgins, VP of marketing at NEI. “Our customers range from startups to Fortune 500 enterprises, and each benefits from our engineering expertise and manufacturing capabilities to solve the numerous challenges involved in getting an application to market.”

NEI manufactures WHIPTAIL’s full range of all-flash storage arrays, including the ACCELA and the hyperscale INVICTA for customers worldwide.

The post WHIPTAIL Taps NEI to Handle Manufacturing of All-Flash Arrays appeared first on WHIPTAIL.

With WHIPTAIL, business-critical applications will benefit from exceptional performance, whether data access is random or sequential and whether IO sensitivity is upon read or write. Our silicon storage array performance metrics are orders of magnitude (100x) above spinning disks to significantly improve database load performance. Varied connectivity options ensure that our high performance storage is correctly plugged in to the overall solution.

WHIPTAIL’s storage offerings are designed and built to efficiently handle the high IO profile demand of critical database applications at a fraction of the cost of mainstream storage vendors.

Read the Database Load Lab Notes.

The post Database Load appeared first on WHIPTAIL.

Flash storage continues to gain buzz in the channel community. Channel consultant Bruce Hockin described, “Over the next three years the enterprise market will witness a seismic shift as flash technology matures from a tier 0/1, optimisation play to a bulk storage alternative.” Read No Room at the Inn? and learn about the distinct advantage WHIPTAIL can offer VADs.

The post No Room at the Inn? appeared first on WHIPTAIL.

The post WHIPTAIL, Blue Turtle Partner appeared first on WHIPTAIL.

The post WHIPTAIL Lunch and Learn appeared first on WHIPTAIL.

The post WHIPTAIL Lunch and Learn appeared first on WHIPTAIL.

To embed this infographic on your site:
<a href=”http://whiptail.com/blog/household-appliances-vs-flash-storage-wattage-reference-infographic/”><img src=”http://whiptail.com/wp-content/uploads/2013/01/Household-Appliances-vs.-Flash-Storage-Wattage-Reference.png” alt=”Household Appliances vs. Flash Storage Wattage Reference – Infographic” title=”Household Appliances vs. Flash Storage Wattage Reference – Infographic” width=”825″ height=”1954″ /></a><br/><a href=”http://whiptail.com”>WHIPTAIL – Performance Storage</a>

The post Infographic: Household Appliances vs. Flash Storage Wattage Reference appeared first on WHIPTAIL.

Achieving Speed and Endurance Out of SSD
Darren Williams, SE Director EMEA & APAC, will be speaking on January 29 at 18:15. He’ll look at the difference between SLC flash and MLC flash and how WHIPTAIL’s unique RACERUNNER software can overcome some of the traditional limitations associated with flash, whilst accentuating the advantages to turn SSD storage into a valuable storage asset for your environment.

An Apple TV will be given away after the presentation to one attendee.

The post Cisco Live UK appeared first on WHIPTAIL.

I have a generator and was able to get most of the house back up and running without issue. With a one week old newborn and a two and a half year old toddler, the generator was a huge relief! I’ve used it a handful of times in the last few years and a couple of weeks ago I had some time to do preventive maintenance. There I was, paging through its manual looking up the maintenance schedules when I stumbled upon a Wattage Reference Guide. It listed the typical wattage draw of household appliances so you could plan for how many appliances could be run safely in the house using the generator. Then, those numbers put a few things into perspective for me with respect to our customers.

Now, I do power calculations for customers all the time. I have for years, for dozens of different environments placed in data centers. Everyone is concerned about the power draw of new components in relation to what the data center is capable of – and rightfully so, power is expensive and finite! But very seldom do I put those numbers in the context of my household appliances. I’ve never thought about how much my house draws versus the data center components I work with every day. That’s why this reference guide was so interesting, and here are a few examples:

The last three have asterisks because the guide mentions these types of appliances may draw up to three times the running wattage upon startup (it makes sense; think of the carousel at the playground when you were a kid – it’s easier to keep something spinning rather than to get it spinning). But that’s important, especially when you only have a finite amount of wattage to work with. In the case of my generator, it’s 5000 running watts, and up to 6500 peak watts. It’s critical that I decide how I use that power most efficiently for my house. Sure, I can get my TV, FiOS box and wireless router, microwave and a few other creature comforts up and running at the same time – but I may be giving up my furnace or the well pump for running water.

What’s that got to do with storage, WHIPTAI, or the price of eggs? Actually, quite a bit (okay, maybe the eggs are a stretch). Data centers do not have infinite power. In fact, these days it’s exactly the opposite. You can only get so much power to a building. Many customers talk to us about how they put a lot of effort into prioritizing those power needs. Now, if your applications suffer from performance shortcomings you need to put some horsepower behind them. The name of the game is IOPS or IO per second. In a traditional storage array, that horsepower will cost you some power draw – it takes power to spin those drives up and keep them spinning. Consider each traditional performance oriented 10k or 15k hard drive will draw on the order of about 10 watts, netting about 180 IOPS per drive. So, 250 drives will net you 45,000 IOPS and will cost you about 2500 watts. Add in a couple of array controllers likely drawing about 400 watts each. That gives us a pretty realistic scenario (and probably even generous) of 14 IOPS per watt.

Our INVICTA array delivers about 650,000 IOPS drawing 1960 watts (just a bit more than a coffee maker). That nets us 332 IOPS per watt – 23 times better than the performance-to-power ratio of traditional spinning disk. Our ACCELA array delivers 250,000 IOPS while drawing 180 watts (three standard light bulbs, based on my handy generator chart). That will net us 1389 IOPS per watt – 99 times what traditional storage is capable of.

Solid-state storage has already begun to radically change how we architect performance for applications in the data center. The data center will forever live within power constraints, just as your home generator will only be able to deliver so much. The question will always be how efficiently you can use that power.

Fortunately, the performance-to-power ratios we deliver make that choice very obvious and extremely compelling.

The post The Power of Solid State Storage appeared first on WHIPTAIL.

Enter WHIPTAIL with 100% silicon storage arrays whose performance metrics are orders of magnitude (100x) above spinning disks to improve the e-mail user experience dramatically. Our varied connectivity options ensure that our high-performance storage is correctly plugged in to the overall e-mail solution. WHIPTAIL’s storage products are designed and built to efficiently handle the high IO profile demand of critical e-mail servers at a fraction of the cost and complexity of mainstream storage solutions.

The post Email appeared first on WHIPTAIL.

Kent Skovgaard will be presenting on Thursday at 11:05 and Friday at 12:25 on the Commaxx stand.

Learn more about the event.

The post Nordic Infrastructure Conference appeared first on WHIPTAIL.

The post WHIPTAIL’s Disruptive Flash Solution Fares Well in the Enterprise appeared first on WHIPTAIL.