Whiptail promises four million IOPS all-flash array performance

Whiptail promises four million IOPS all-flash array performance

All-flash array maker Whiptail has announced an upgrade to its Racerunner operating system that will allow its Invicta NAND flash arrays to scale to 30 nodes and up to 360TB of storage.

The all-flash array company claims that scaling to such capacities could bring up to four million IOPS (input/output operations per second) and throughput of around 40GB per second.


Currently such levels of performance represent an extreme that is only required by a handful of organisations worldwide. But Whiptail is aiming its new capabilities at organisations that want to move all their high-performance storage requirements onto a single array.

Max Riggsbee, head of products and marketing at Whiptail, said the platform is designed for multi-tenant storage requirements.

“Where you might see organisations buy a different array for different needs – say Oracle on one, VDI [virtual desktop infrastructure] on another, batch processing on another, none of which needs one million IOPS – we suggest they could consolidate onto one array with a smaller footprint and not worry about performance.”

The Invicta Silicon Storage Array is a NAND MLC flash-based all-flash unified storage (file access plus Fibre Channel and iSCSI) array product.

It comprises Whiptail Silicon Storage Routers that provide some processing power, connections to networks and other storage nodes, and some features such as RAID management, replication and snapshots. Meanwhile, Silicon Storage Nodes comprise capacity and storage controller with functions such as RAID and snapshotting on board.

Previously, Invicta scaled to six nodes. The announcement marks an upgrade to Racerunner version 4.2 that will see it scale to 30 nodes. Routers and nodes can be mixed and matched to scale capacity or performance.

All-flash arrays have risen in prominence over the past year or two to meet a need for very fast processing, in use cases such as financial transactions and for virtual desktop environments.

Initially, start-ups such as Whiptail and Violin Memory blazed the trail, but more recently the big storage suppliers have taken notice and made acquisitions, such as EMC’s purchase of XtremIO, IBM’s buy-out of Texas Memory Systems and Hitachi Data Systems’ addition of an all-flash module to its VSP enterprise arrays.

Riggsbee gave examples of organisations Whiptail is aiming the product at. They include financial services for batch processing, transaction processing, market data and analytics; customers that want to boost website response times; file conversion in the media industry; and CAD drawing processing in construction.

Read More

Charityshare gets Violin flash array for 2,000 seat VDI rollout

Charityshare gets Violin flash array for 2,000 seat VDI rollout

IT consortium Charityshare has deployed a Violin Memory 3000 Series all-flash array to support the roll-out of up to 2,000 virtual desktops for its Age UK operations.

Violin was chosen on grounds of cost and the company’s backing from HP over Whiptail and Texas Memory Systems (TMS). An all-flash array approach was chosen over mixed flash and spinning disk because of concerns that automated storage tiering features in the IBM V7000 array would not provide the availability required for virtual desktop infrastructure (VDI) operations.


Charityshare provides enterprise level IT to three charities, Age UK, The Alzheimer’s Society, and The Children’s Society from two datacentres, in west London and the Docklands.

Its existing storage infrastructure comprised HP EVA and Hitachi Data Systems AMS 2300 SANs but this was deemed unsuitable for use with the Age UK virtual desktop roll-out due to issues of contention for I/O suffered by the other charities on the HP SAN.

Greg Cooke, engineering manager with Charityshare said, “We had been running VDI – Windows XP on XenDesktop on VMware – for two or three years for The Alzheimer’s Society, and The Children’s Society from the EVA SAN. But, we had Exchange and SQL on the same SAN, and there would be issues with, for example, SQL taking all the I/O, so we decided we needed dedicated storage for VDI.”

Cooke’s team calculated it needed between 10 and 80 IOPS (Input/output operations per second) per user and averaged that to 35 IOPS per user, with a 5GB Microsoft Windows 7 virtual desktop. Initial deployment was to 800 users with capacity to expand to 2,000 when the other charities move to the same VDI infrastructure, which would require 70,000 IOPS.

After evaluating options including mixing flash and hard drives in a SAN and other all-flash array products, Charityshare went for a Violin Memory 3000 Series array with 12TB of MLC flash Violin Intelligent Memory Modules (VIMMs).

VIMMs are Violin’s proprietary storage media, sourced via agreement with Toshiba and built onto cards with Violin software on board to provide data protection and other functions. They are hot swappable, with data failing over to spares on VIMM failure.

The 3000 Series is sold as direct-attached storage that connect to a physical server via PCIe in contrast to Violin’s 6000 Series which are marketed as SAN arrays that provide shared storage via a fabric.

Other flash suppliers evaluated following recommendations from Gartner were Texas Memory Systems (now owned by IBM) and Whiptail. TMS was ruled out as too costly. Violin was chosen over Whiptail because of its backing from HP. This was seen as important by the Charityshare IT team because all-flash arrays are a new technology and the organisation wanted guarantees from a big storage supplier.

Other options examined included IBM’s V7000 enterprise SAN with a flash tier.

Cooke said this was ruled out after the organisation heard of the "Monday morning problem", where the automated tiering algorithm moves data off flash to lower tier disk over the weekend because it isn’t being used very much.

“Sure, it would reconfigure it to move that data back to flash but VDI needs its data from fast storage straight away,” he said.

All-flash arrays have risen to prominence in the last year or two, initially from startups such as Violin, Whiptail, Pure Storage and Kaminario, and have more recently gained big supplier backing with EMC’s acquisition of XtremIO and IBM’s purchase of Texas Memory Systems.

Read More

Nutanix supports Red Hat KVM in upgraded storage stack hardware and OS

Nutanix supports Red Hat KVM in upgraded storage stack hardware and OS

Converged storage and server maker Nutanix has announced upgraded hardware and a new operating system (OS) version with support for Red Hat's KVM hypervisor in addition to existing VMware support.

Nutanix debuted in 2011 with its NX-2000 Complete Cluster storage stack aimed at server and desktop virtualisation use cases. This comprised 2U blocks with Intel Westmere processors, plus PCIe and SATA flash drives and hard disk drives (HDD) with the VMware hypervisor built in. Marketed as a "datacentre in a box", the product could be scaled by adding compute and storage capacity.


Nutanix has now added the NX-3000 hardware format, which sees processing power boosted with an upgrade to 8-core Intel Sandy Bridge chips, increased memory and two instead of one 10 Gbps Ethernet ports per node.

The OS upgrade to version 3.0 includes replication to other Nutanix instances and disaster recovery policy configuration that includes the ability to specify data protection domains with combinations of primary and secondary sites.

There is also high-availability failover and failback functionality, support for another hypervisor – Red Hat’s KVM – and compression of blocks of data according to usage patterns.

Red Hat KVM is a hosted hypervisor, which means it is more efficient than the likes of VMware and Microsoft’s Hyper-V. KVM has a tiny share of the virtualisation market at present, but is reportedly used by some big cloud users, including Google, IBM and Rackspace. Nutanix cites Hadoop as a use case for its KVM variant arrays.

The latest Nutanix OS also has auto-detection of new Nutanix nodes, which means they can be added without downtime and the OS will auto-balance data across the newly expanded pool.

Read More

Top 10 server virtualisation stories of 2012

Top 10 server virtualisation stories of 2012

The benefits of server virtualisation are indeed quite impressive for today's server and datacentre administrators, but what about you – the storage manager? While everyone else is reaping the rewards of their newly virtualised servers, you are realising that your job just became a lot more complex and time-consuming.

We feel your pain. For that reason, Computer Weekly has compiled a top 10 list of tips and guides to server virtualisation to help you address the variety of storage-related challenges you may face in your virtual server environment.


Best practices for using server virtualisation in your storage environment
Server virtualisation can lower costs and increase capacity, but there are a number of best practices to consider when using or thinking about server virtualisation. Check out this tip to learn four key things to keep in mind when using server virtualisation in your storage environment.

Virtual server and virtual machine backup: Hot data backup storage technologies

Virtual machine (VM) backup is still in its adolescence, but maturing fast, with significant developments that should offer some relief to beleaguered backup administrators. Explore this report to learn more about the basics of VM backup.

iSCSI versus Fibre Channel in virtual server environments
Deciding which SAN technology to use for your virtual server environment requires careful consideration. Find out for yourself by learning the pros and cons of iSCSI versus Fibre Channel in virtual server environments.

Everything you need to know about vSphere and data storage
System administrators are still gushing over the benefits of VMware's virtualised servers, but for storage professionals, the shift in server architecture has caused new headaches. Find out how VMware vSphere aims to address many of these storage-related shortcomings.

Storage virtualisation benefits: Increasing flexibility and utilisation with virtualisation
Our e-guide will help storage professionals understand storage virtualisation benefits and get started with their own storage virtualisation project by providing purchasing tips, file virtualisation best practices and more.

I/O virtualisation: Approaches and implementations
Learn which companies have offerings in the I/O virtualisation space, how they approach the network infrastructure bottleneck problem, what the pros and cons of each product are, and the differences between I/O virtualisation and Fibre Channel over Ethernet (FCoE).

Including server virtualisation in your disaster recovery plan
This tip discusses best practices for including server virtualisation in your disaster recovery (DR) plan based on the hypervisor in use, the IT infrastructure in place, and the required recovery time objective (RTO) and recovery point objective (RPO).

How to buy a blade server
Blade servers can make your datacentre more efficient, but a blade project can be costly. In this expert tip, we tell you what questions you should ask before committing to a blade server supplier.

Storage virtualisation essentials: Cut costs and increase storage efficiency
Storage virtualisation can cut equipment costs and increase storage efficiency. Explore this guide to discover the pros and cons of storage virtualisation architectures and learn about storage virtualisation products and how storage virtualisation fits into disaster recovery plans.

Guidelines for implementing virtualisation in your storage infrastructure
Virtualisation has shown up in all layers of the storage infrastructure and storage managers need to balance the benefits of virtualisation against the complexity it brings. Here are some guidelines for implementing virtualisation in your storage infrastructure.

Read More