In part one I looked at the overall macro trends in the desktop virtualization market, now in part two I want to look at what to expect from key vendors and vendors: Microsoft, Citrix, VMware, and AppSense as well as product groups such as thing client and storage vendors. All with an eye to Desktop Virtualization in 2012.
Aside from its unwavering commitment not to extend support for Windows XP beyond the April 2014 deadline, Microsoft delivered little to encourage desktop virtualization adoption in 2011, and I do not expect there to be any significant change in 2012 either. The much-maligned “VDI Tax” (Windows VDA license) that is Microsoft’s $100 per year VDI license for non-Windows endpoints is unlikely to change. The best that Microsoft will do here will be to tweak the rules for thin client and mobile endpoints to be more favorable towards Windows 8 on ARM processors, perhaps eliminating the VDA license requirement for this class of devices in favor of a conventional Software Assurance licensing model.
Citrix has to continue with its the steps it took in 2010 to both broaden its services, and simplify the adoption of desktop virtualization. Citrix is too invested in VDI to allow it to fail, so must take a leading role in simplifying and reducing risk in the VDI market as I outlined in Part I. At the same time to accommodate the needs of customers who do not need VDI in any form, Citrix must provide a standalone, easy to use, easy to buy, easy to understand, alternative to the distributed desktop virtualization offerings from Virtual Computer, and MokaFive. XenClient as component of XenDesktop is not the answer.
Citrix’s wild card is the old Kaviza team. When Citrix acquired Kaviza in May 2011, Citrix CEO Mark Templeton made it clear that for the time being Kaviza would not the integrated into the core desktop virtualization group but would remain as a semi-independent entity. So far, Citrix has remained true to this vision, and has recently indicated that the Kaviza team will be transitioning into a dedicated SMB unit. With this in mind, and assuming that the Kaviza team is allowed/encouraged to retain the same autonomy and willingness to compete, I would not be surprised if the
Kaviza SMB team don’t have their eyes on some sort of partnership with Pano Logic as a way to extend VDI-in-a-Box into a comprehensive single SKU solution. Admittedly, this is a bit of a stretch as it would put Citrix into competition with the many thin client vendors that it partners with today, nevertheless I don’t think it beyond the bounds of possibility that Citrix and Pano Logic will establish a closer relationship in 2012.
VMware’s End User Computing team suffered a set back in November 2011 when Chris Young left to head up Cisco’s Security group. Young has been replaced by Boaz Chalamish who moved over from VMware’s Management Business Unit. It’s not yet clear if Chalamish intends to change the direction that Young set in motion but regardless of this there are certain key actions that were set in motion in 2011 that must be completed in 2012.
VMware’s vSphere 5 launch was marred by the late cancellation of ”View Accelerator” - officially Content-Based Read Cache (CBRC). This feature would have improved the performance of VMware View by reducing SAN reads of sharable virtual machine images, but the management interface which was to be delivered as part of View was not ready, so even though the CBRC code was present in ESXi 5 the announcement was pulled at the very last minute. VMware must address this loose thread and deliver a working View Accelerator early in 2012 if it is to retain the momentum it gained with View in 2011. The other mandatory deliverable is a PC-over-IP client for Apple Mac OS X. This has been missing in action for too long and needs to be addressed if VMware is to retain credibility as an enterprise desktop virtualization vendor.
After achieving a massive buzz with the announcement of Project AppBlast at vmWorld there has been little news from VMware as to progress. Considering the level of positive feedback it received, AppBlast has the potential to give VMware’s end user computing team a big boost, but it has to deliver. A beta early in Q1 2012 would be a strong sign that VMware is progressing to a v 1.0 release in Q2.
Most importantly though VMware needs to follow Citrix and Quest’s lead; it must look to the gaps in its current services portfolio and plug them through acquisition.
So far AppSense has done little to justify Goldman Sachs’ $70 million investment beyond releasing a private beta of the Strata user installed application management technology enabler. AppSense has already committed to delivering a public beta of Strata in Q1 2012,but must deliver a rock solid v 1.0 product no later than May to coincide with the Citrix Synergy conference in San Francisco if it want to confirm Goldman Sachs’ assessment. Even so this will not be enough, 2012 has to be a breakout year for AppSense. Most importantly AppSense has to start laying the foundations of its future public incarnation (AppSense has indicated that it hopes to IPO in 2013/2014). This means that it must move beyond the narrow set of products that currently defines it and establish a border footprint. Given the strength of the personalization and policy engine at the heart of AppSense software, it would be a logical step for it to move into the post-PC solutions space and extend its capabilities to personalizing Web and mobile applications. It has already shown a willingness to partner with UK unified end-user computing software specialists Centrix Software and could well benefit from closer ties in the future. Whatever it does, to ensure its long-term growth, AppSense must reinvent itself. In the same way that Citrix successfully transformed itself with harming its relationship with Microsoft in 2011; so in 2012 AppSense must reinvent itself without damaging its relationship with Citrix.
Thin client vendors
One of the more surprising announcements at Citrix Synergy Barcelona was the announcement of the “Citrix HDX Ready System-on-Chip” initiative. Even more surprising, was the announcement that that NComputing was a launch partner for the new initiative. NComputing dominates many emerging markets with its low-cost (and relatively low performance) thin client system. Given the very low cost of NComputing devices, it is unlikely that NComputing will look to HDX on a chip here. Instead it is more likely to use this technology as a means of establishing a beachhead in enterprise IT organizations that until now have largely been served by more expensive solutions from HP, Dell, Wyse Technology etc.
HDX on a chip will serve to level the playing field between thin client vendors, making product differentiation harder. Assuming that HP does not reverse itself again and decide after all to get out of the PC business, there is still value in HP in being able to offer its own thin client hardware, even if all it do is to repackage hardware designed and manufactured by someone else. The same rationale also applies to Dell, IBM and perhaps Lenovo. Similarly Wyse Technology is unlikely to suffer. Aside from its strength as a thin client vendor, Wyse has already started to broaden its portfolio, the recent acquisition of Trellia will significantly bolster its position as a provider of endpoint management services, and it may well continue down this path, looking at other mobile technologies as a way to further diversify.
One company that may find itself under pressure is Pano Logic. Today, Pano Logic’s systems are competitively priced, however if the predictions of high-performance thin client hardware dropping to $100 – $150 are fully realized, Pano Logic could face a degree of pricing pressure. If prices remain above $150, Pano Logic should not have any problems. However, if a high-performance HDX thin client was available for $100 Pano Logic could well find itself struggling. Having said that, Pano Logic has some room for maneuver in both manufacturing costs and product pricing, and its products are uniquely easy to use which does provide it with a degree of protection. Far more at risk, are the second and third tier thin client vendors that have little to differentiate themselves and will see the price differential between the solutions that they can offer and those of the top-tier vendors eroded by the falling prices. Under these circumstances, I feel it is inevitable that one or more thin client vendor will have fallen by the wayside before 2012 is over.
2012 should be a good year for storage innovators such as Virsto, Atlantis Computing, Fusion-I/O etc. The rising price of hard disk storage brought about by the flooding in Thailand in October and the falling price of flash storage will have a positive impact on all these companies. The only uncertainty is whether or not these companies will survive the year as independent entities or if they will find themselves subject to acquisition.
There are clear customer benefits to be had from getting the advanced storage of the migration capabilities that Virsto and Atlantis Computing offer built-in to the hypervisor, and at a time when competition between hypervisor vendors is hotting up, baked this capability in to the hypervisor could be a worthwhile differentiating factor. Even so, it’s hard to judge how this might play out, Citrix already offers a limited storage optimization capability in XenServer Intellicache, and VMware pulled the announcement of its own “View Accelerator” from the vSphere 5 launch, even though the code is still present in ESXi 5, so it is a clear if either would consider the benefits of acquiring a Virsto or an Atlantis to be worth the investment. Microsoft however might see things differently; Hyper-V lacks this capability and Microsoft is clearly committed to offering a competitive hypervisor platform.
The higher cost of hard drives will mean that SAN specialists will either be forced to accept reduced profit margins or pass on the increase in hard drive costs direct to their customers, who will inevitably then look more closely at additional third-party storage optimization solutions. Which will inevitably drive storage vendors to look more closely at making acquisitions in this space. Regardless of which path they take, the attractiveness of any technology that can reduce the overall cost of storage by better aligning throughput (IOPS) with capacity will significantly increase in 2012.