Chris Wolf, Burton Group senior analyst, analyzed Citrix’s acquisition of XenSource in a recent Burton Group blog post. He sizes up the situation, saying:
“While having the technology is one thing, bringing it to market is an entirely separate issue. This is where the Citrix acquisition makes great sense for XenSource. Financially fueled by Citrix, XenSource now has the financial clout, sales, and channel resources to go after the large stake of unclaimed virtualization market share in the enterprise. Don’t get me wrong. This will not be easy, as Citrix and XenSource are competing against powerhouse vendors with strong sales, channel, and IHV partnerships. VMware, Microsoft, Red Hat, and Novell are well established in the enterprise, and are all looking to add to their share of the market. Virtual Iron has been making a lot of noise in the SMB space lately, and they should see the explosion of the XenSource sales channel as a serious threat.”
Wolf sees the acquisition as a win for Citrix and Xen and for users, too.
“In the coming months and years, we should expect to get enterprise-class virtualization technologies at lower costs, with more features, and a motivated group of vendors that are eager to push innovation to remain competitive.”
Read his blog in its entirety on the Burton Group Data Center Strategies blog.
XenSource is now going to be a part of Citrix, an interesting move that answers a question that a lot of Citrix aficionados, myself included, were wondering – “What is Citrix doing to remain relevant in the face of Virtual Desktops?”. Virtual Desktops are what this is really all about. Citrix is not a server company, it’s a delivery company. In terms of quality and speed, you can make the comparison that Citrix is the FedEx of bits and bytes, if you will, to Micrsoft’s US Post Office. Citrix, with its ICA protocol, is like FedEx in that it is faster, more full-featured, and more expensive than the trustworthy old Post Office of RDP. Citrix may make a line of server products, but their business is in getting applications into the hands of end-users regardless of distance, client operating system, link speed, and other factors that typically inhibit deployment. Whether this is done by granting remote access to applications over the web, over a proprietary protocol, or any other method has been irrelevant to Citrix for some time – they adapted quickly to the web, enabling ICA-delivered application delivery as early as MetaFrame 1.8 in the late 90’s. What was always a problem in the Citrix world was the full desktop – Citrix’s own CCA (Citrix Certified Administrator, a cert I hold near and dear) training pushes for application delivery over desktop delivery, and for good reason. Presentation Server, the successor to MetaFrame (which was the successor to WinFrame) runs on a Windows server, and there are as many inherent risks of letting people have remote desktops on a server as one can imagine, not the least of which is registry corruption and accidentally deleted files that can hose the entire server, as well as disk space issues. Imagine 50 users saving the same 200MB PDF or PPT of the company’s end-of-fiscal-year financial report on their desktop instead of a network file directory, and you can see how the problem would start. Couple in the complexities of antivirus / antimalware and you can picture the resource spikes every morning at 9am when 20 users log on and the AV program starts running it’s at-logon full computer scan 20 times. While some AV products are Citrix / Terminal Services aware and can be configured not to do this, that’s the exception, not the rule – most often the schedules need to be changed from the default of daily scanning on logon that many companies use and adjusted to only a once-daily at a fixed (off-hours) time. Then of course there’s the resource issue – one errant application that cpu-locks the box takes it away from everyone, because it’s all really just one computer. While Citrix has made many inroads to mitigating these problems, they still exist, and make full-featured Citrix desktops hard to manage and hard to justify on a large scale.
What XenSource adds to the Citrix portfolio is a virtual desktop platform without the caveats of the delivered desktop being part of the same server OS as the application – each desktop can be shunted off onto shared storage, is it’s own contained OS, and can be delivered seamlessly without other users vying for resources. There are still resource issues to contend with such as CPU locking, but because virtualization allows for greater restrictions on access to the physical hardware, these can mitigated faster and without large-scale interruption. The same applies to memory locking, but to an even greater extent – since a VM has X-amount of memory, you can lock it all up and not affect the other virtual machines much at all. No doubt Citrix is working on perfecting desktop delivery of Xen-based desktops over ICA. Likewise Citrix is no doubt working on integrating their already-impressive management tools (MetaFrame XPe even had a configurable chargeback function built in from it’s initial release!) to make managing Xen desktops easier. What’s further is that this gives Citrix a hedge against the possibility of Linux desktops making significant inroads into the enterprise market in the next few years. Instead of depending on Windows, Citrix has just become open to delivering multiple flavors of Linux, BSD, and many other operating systems to users who need them.
What about running Citrix on Windows on Xen host servers? I don’t think Citrix cares one way or the other if you do – it’s still a license for the software regardless of whether the host is virtual or physical, so other than a marketing play about getting scalability on virtual hardware, there’s not much advantage to be seen. Thin Clients? Citrix doesn’t sell them, but they have very cozy relations with all the TC vendors out there, especially Wyse. Wyse has already made great inroads with VMware virtual desktops, and it’s sure to follow that they’ll support Xen desktops as well.
What if I’m wrong? It’s a good possibility. And I’m sure that Citrix sees the other side of the coin – the death of the full-featured desktop as we know it, with a glorified web browser supplanting it. Indeed, this would be the ideal Semantic Web / Web 2.0 client operating system – all applications occurring in a browser, with nothing on the user’s side that can really break, or take a long time rebooting, or need a lot of skill to repair, or require complex driver slipstreaming into images, etc. etc. Just a browser and some horsepower to help the client-side of the web-applications get processed is all anyone would need for an all-web-application environment. This is basically just a distributed thin client, with a slimmed down OS and remote application delivery. This just happens to also be the business Citrix has been supporting for decades. Many of these applications, so thin to the end-user, are fairly large on the server side, and being able to provide virtualization of those apps will keep Citrix relevant when the user’s choice of client no longer matters because everything is designed for web delivery. Right now that’s a selling point for Citrix – they have an ICA client for almost every OS out there, giving people on Linux access to Windows apps without messing with WINE, giving remote users with limited bandwidth access to giant client-server applications. They have web-enabled gateways to allow this to happen over a browser (with a plugin). The trouble with ICA is that if the need for specialized software to connect to remote systems diminishes as more and more applications because browser-based, the need for such a system dependent on plugins and other protocols wanes, leaving Citrix’s ICA out in the lurch. So, either way the cookie crumbles, Citrix has a smart, effective strategy for maintaining relevence.
There are many questions unanswered, such as what will happen to future Xen development? Will Citrix close any new Xen code or keep it Open Source? What will happen to Virtual Iron? Will Xen fork? What will happen to the long-standing but sometimes strained relationship between Citrix and Microsoft? Stayed tuned… this is an interesting time for virtualization and Citrix followers, and I would expect a lot more blogging, reporting, and gossiping about what this means in the very near future. IT Managers should keep an eye on their Virtual Desktop initiatives, for sure. Overall, I would expect this to be a very positive thing for virtualization on the server, desktop, and application front. Disruptive, no doubt, but positive.
Alas, another kernel release and another VMware breakage. Sadly, staying on the cutting edge of kernels bites you sometimes. This is one of those times.
As detailed here:
No Internet on VMware WS w/ Linux 2.6.22 kernel
Kernels before 2.6.22 are able to use VMware networking, but due to kernel changes, at this moment, VMware’s networking isn’t reliable. Even attempts to fix, like bridging, did not work. Even patching with any-any-112 patches haven’t worked.
Not a big deal, as I’m sure it’ll be patched any day, but *this* day, your author will be running an older kernel.
Linux can match and outshine Windows in data center — thanks to its strong security, reliability, and long term usability — but it has upsides and downsides on the virtualization side, according to software engineer and Linux and VMware Server technologist Chris Berg. In this excerpt of our Q&A, Berg gives a power-user’s view of Linux’s strengths and weaknesses on the virtualization front.
“By providing a solid, low or no-cost platform out of the box, strong virtualization solutions abound on Linux,” Berg said. “Solutions such as Xen or VMWare ESX server put Linux on even ground…by offering a comprehensive virtualization solutions and easy to use interface.”
My interview with Berg is an offshoot of coverage of the recent LinuxWorld/Next Generation Data Center Conference. Getting connected with veteran users is the best thing about covering these shows, I think. So, here’s what Berg had to say.
In what ways will virtualization spur wider adoption of Linux on the server side?
Berg: A variety of both open and proprietary solutions are available for Linux including KVM, Xen, VMWare ESX Server. The cost of these solutions ranges from nothing up to tens of thousands of dollars. The benefit to a wide variety of virtualization solutions is that an organization is free to choose a solution that fits their needs. The variety of choices and costs is what will spur the adoption of Linux on the server side.
Whereas some Linux virtualization solutions are not “bare metal”, installed on an existing Linux server products, others such as VMWare ESX server consist of a fully-installed Linux distribution under the hood. This enables the ESX server to be maintained like an appliance, with little operating system level interaction.
In what ways will virtualization make no difference in adoption of Linux on servers?
Berg: The current virtualization trend does not translate to across the board adoption of Linux. In cases where a virtualization solution isn’t “bare metal,” there won’t be as large a push for Linux due to internal support staff’s familiarity with Windows servers. Due to the variety of solutions available for virtualization on Linux, there can be a variety of interfaces to maintaining an installed solution. This can be daunting for IT support staff and those seeking third-party technical support.
What technology or market trends will help increase Linux server adoption in the enterprise in the rest of 2007 and early 2008?
Berg: As the virtualization push continues, Linux server adoption in the enterprise will continue to increase. Either through a solution that relies on an installed Linux or a solution like VMWare ESX Server that uses Linux under the hood.
The continued push for higher value offerings will see Linux continue to be deployed in essential services roles, such as providing DNS, DHCP, proxy services and continuing into the application server space as many independent software vendors continue to invest in Open Source.
In general, what barriers stand in the way of adoption?
Berg: One of the largest barriers to Linux server adoption in the enterprise is the back room nature of Linux success stories. In many large enterprises Linux is deployed on the server and performs capably, but the word never gets out.
Help SearchEnterpriseLinux.com spread the word about Linux and virtualization by telling us about your deployments. Comment here, or write to me at firstname.lastname@example.org.
VMware finally staged its initial public offering (IPO) today, with EMC selling 10% of its stake in the company, or 33 million shares, and raising almost a $1 billion in the process. The shares were priced at $29 – higher than expected just last week – but by midday had shot up to over $50 per share.
To anyone who’s watched the company over the past couple of years, that the VMware IPO was a rollicking success should come as no surprise. Quarter upon quarter of consecutive growth. Tremendous goodwill from its partners and customers. A slew of imitators and competitors. A healthy ecosystem or start-ups hoping to capitalize on its success. Even entire websites devoted to the core VMware technology
More interesting, I think, is how VMware’s newfound semi-autonomous status will impact it product development, and more importantly, how the company deals with its customers?
Public companies are different from private companies, which are different from wholly owned-subsidiaries (even independently operated ones). Sure, right now, VMware can do no wrong, but how will having Wall Street breathing down its neck every quarter impact the company’s ability to create important and fully baked software? How will the emphasis on quarterly results influence the VMware sales force, and the way it treats its accounts? Will VMware’s status as a publicly traded company affect its relationships with its partners?
I don’t know the answers to these questions, but will be looking out for clues over the coming months. In the meantime, if you have any theories, feel free to leave a comment, or shoot me an email.
Check out the Enterprise Linux Log for news about Asigra’s new 64-bit agentless backup for virtual servers.
Virtual Iron’s Mike Grandinetti provides insights about the synergies between virtualization, blades, server consolidation and iSCSI in this interview with Jan Stafford, SearchServerVirtualization.com’s senior site editor.
Bernard Golden discusses some take aways from Tony Iam’s session at the Next Generation Data Center conference in San Francisco.
In this video, senior site editor Jan Stafford discusses the fact that KVM, Xen and other virtualization platforms are not interoperable, a subject that troubles users and was given short shrift by vendors at LinuxWorld/Next Generation Data Center Conference 2007 sessions.
Xen expert Bernard Golden sounds off on Microsoft’s presentation at LinuxWorld 2007 and the impact of Novell’s new Linux SUSE and virtualization products. Golden is a systems integrator, SearchServerVirtualization.com expert and author of the upcoming book, “Virtualization for Dummies”.