Flash changes everything. As Wikibon has been discussing for a few years now, flash has a significant impact on system and storage architecture. For a long time, storage consumption was measured in capacity and cost ($/GB). With the advent of flash, performance and latency have become very important (see Storage Cost as a Function of IOPS).
Since latency between flash and the application is critical, there has been a move to bring storage closer to the compute (either in or very close to the server). Cisco has been pushing deep into data center architectures with its UCS product line and today announced the intention to acquire WHIPTAIL for approximately $415M in cash and incentives. As interesting as the acquisition is that WHIPTAIL, which offers an all-flash array, is positioned as part of the UCS (server) portfolio.
While this is not a declaration of war against its storage partners, Cisco’s acquisition will add to competitive tension as the boundaries between layers of the stack are blurred and account teams fight for the same IT dollars.
Cisco in the storage market
When Cisco launched the UCS product line, it sent a ripple through the enterprise IT vendor ecosystem. HP bought 3COM, IBM bought BNT, Dell bought Force10 and while there has been a realignment of the networking industry, Cisco has remained dominant.
Cisco has taken the leadership position in driving the trend of converged infrastructure. In addition to the well documented VCE (VMware, Cisco, EMC) joint venture, NetApp/Cisco FlexPod solution and Cisco’s participation in VSPEX reference architecture, Cisco has expanded its storage partnerships to include Nimble Storage, Tegile, Tintri and many others. Cisco is heavily reliant on storage partners to help sell UCS solutions.
Cisco conservatively has billions of dollars of sales related to the storage partnerships this year (not just through the converged solutions, but also through joint engagements). For many years, it had been rumored that Cisco might buy EMC or NetApp, but even with its large cash reserves, either of these acquisitions would be prohibitively large.
- Flash = new opportunity
Flash offers a new opportunity to win a greater influence and share of IT spend compared to traditional arrays. WHIPTAIL makes all-flash arrays (see my Breaking Analysis from earlier this year – full video below) and these solutions are designed for the highest performance and lowest latency applications, which means that the solution must be pulled into the “server network” (low latency Ethernet solutions like RoCE are a high likelihood for Cisco environments).
Even before the acquisition, Cisco’s UCS solutions offered server-based flash – from EMC, Fusion-io and others. Cisco has done an excellent job at growing UCS to the #2 position in the bladeserver market (still low single digit market share in the overall server market). Not having an in-house flash solution would soon become a limit to Cisco’s growth.
- Challenge in conflicting storage partners?
The challenge that they have is that there is great risk that conflict with storage partners could cause the loss of more revenue than the WHIPTAIL solutions are likely to create in the next 1-2 years. By positioning the acquisition specifically as part of the UCS, Cisco signals that they want to maintain all of the server/desktop solution partnerships. As EMC (XtremIO) and NetApp (FlashRay) all-flash solutions push into the market, there will be competition, but these are not currently the primary products that are part of the VCE, VSPEX and FlexPod partnerships.
Cisco, EMC, NetApp and the other storage vendors are all scrambling today to respond to this move. From a user standpoint, today’s announcement is just another sign that flash is making big waves (just yesterday, Western Digital bought Virident for $685M).
Customers are rapidly adopting more flash and are looking for strategic vendors to provide these solutions. Customers push vendors to work together when they decide to buy solutions from multiple vendors. I heard from a number of Cisco customers at VMworld that they are strongly considering moving to NSX (check out our videos with Cisco/Insieme and Nicira/VMware from VMworld); despite the saber rattling, Cisco will need to support these customers.
Customers do not turn to Cisco today for storage decisions, so it is smart that Cisco takes a UCS-led direction for positioning the WHIPTAIL acquisition.
Cisco will not abandon any of the current initiatives, rather the acquisition allows for an expansion of its portfolio without directly threatening the vast majority of existing partnerships and engagements. Flash economics continue to improve and customers should be reviewing the impact and options on entire system design (not just the individual compute, network and storage pieces) during each refresh cycle.