We can talk until we’re blue in the face about universal clients, ubiquitous data access and streamlined image management, but ultimately the question of whether virtual desktops make sense comes down to what IT decisions always come down to: money.
Johnathan, a Server Virtualization blog reader, recently posted a comment on one of my posts detailing the math for a 250-seat virtual desktop infrastructure (VDI)/thin-client implementation, which amounted to a $350 per-desktop-capex advantage for VDI; a three-times faster deployment schedule and troubleshooting times that were orders of magnitude faster (albeit harder to quantify). Not too shabby.
Of course, that was before VMware announced new pricing for its re-branded VDI suite, View 3. At $150 per seat for View Enterprise or $250 for View Premier, capex savings would decrease to $300 or $200 per desktop. That’s assuming you pay list price, which is highly doubtful. But it also doesn’t account for the storage capacity savings you might realize by using View Composer to share desktop images: an average of 70%, according to VMware.
Suffice it to say that assigning ROI dollars to an IT project is a highly personal, subjective affair. And that the numbers posted by others are often suspect, as Bernard Golden points out in his article “Virtualization Projections Deserve Scrutiny.” Here, Golden looks into a Butler Group report that reports client virtualization savings of $159,000 for 1,000 desktops, or $159 per desktop, per year. Come to find out, the $159 savings was in energy costs alone. Who knows what the overall cost of the deployment really was?
At any rate, if you’ve done the math on a VDI implementation, and believe that your numbers bear scrutiny, go ahead and post the numbers in the comments section of our blog.
There’s a lot of virtual desktop news these days, and before too much time passes, I want to share some tidbits on VDI that I picked up this week and that had never occurred to me before.
- VDI can save you money on software licenses. At least, that’s what I hear from Jeff Cunningham, a network administrator at the Agricultural and Resource Economics department at the University of Maryland, who implemented about 70 virtual desktops for faculty, staff and graduate students. For instance, an individual license for the data analysis and statistical software package Stata runs about $700. In contrast, a 10-seat network license costs the university $2,000, for a savings of $5,000, and the budget to deliver interesting software to a greater number of students.
- Thin clients can withstand a long power outage. Kunal Patel, the IT director at Nina Plastics, whose VDI project I wrote about earlier this week, told me that during a recent power outage, the company’s regular desktops drained their APC battery backups in less than 10 minutes. Their Pano Logic thin clients, on the other hand, stayed on for four hours. In a similar vein, the University of Maryland’s Cunningham stuck a kilowatt meter on a bank of five Pano devices and a bank of five regular desktops and discovered that the Pano devices consumed one-fourth the power of the regular desktops.
- Some IT managers are skeptical of thin clients’ supposed cost advantages. As an example, check out Basilm’s comments on the Server Virtualization Blog. What about you, dear Server Virtualization Blog readers? Have you done the math on VDI and thin clients? What’s the verdict?
- Big companies need big security. With their strong security and compliance needs, verticals like finance, health care and government are a natural fit for VDI. But in order for them to adopt it, the VDI community needs to support biometric authentication mechanisms, such as fingerprint readers and face recognition software.
That’s all for now, folks. Brace yourself for a lot of news on virtual desktops. Things are about to get interesting 🙂
Have you ever marveled at how fast desktops and laptops start breaking down, even under normal working conditions? Try putting a desktop on the floor of a plastics manufacturing facility. You’ll be lucky if you get a week out of the desktop before something fails, said Kunal Patel, IT director at Nina Plastics USA in Orlando, Fla.
The production facility at Nina Plastics performs a process called plastics extrusion, which releases all manner of dust and grime into the atmosphere, clogging up fans and power supplies, and settling down on hard drives, Patel said.
At first, Patel’s staff would try and fix the broken desktops, which production workers used to log their job start and stop times. “But it became too much of a hassle for IT to constantly fix stuff,” Patel said, so the company eventually gave up on trying to computerize its production facility.
“We shouldn’t be maintenance men,” said Patel, who also oversees application development for the firm. “We all went to college and should be working on more important problems.”
However, that was before Patel, with a handful of administrative staffers, spearheaded a virtual desktop trial using a combination of VMware virtualization plus thin clients from Pano Logic.
By going with virtual desktop infrastructure (VDI), Nina Plastics derived all the usual benefits you’d expect: faster desktop provision, easier patching and upgrading, simplified troubleshooting, etc. At the same time, Patel also found that the Pano Logic devices were robust enough to withstand the harsh conditions of the production floor. “There’s no CPU, no memory, no fan. There’s really nothing in there to break or get old,” he said. The company has since reintroduced computers into its production facility, giving customer service staff real-time visibility into the status of a particular job.
Patel also plans to add touch-screen monitors to the Pano devices, a feature v and supported in the Pano Virtual Desktop Solution (VDS) 2.5 software.
Patel had lots of other interesting stuff to say about his VDI deployment, but for now, suffice to say that he’s a fan. “It’s easy to fall in love with, especially when you have suffered so much,” Patel said. “I have fewer gray hairs, fewer lost girlfriends, and a lot of time given back to me because of virtualization.”
Earlier this month, I wrote about how the PCI standard was recently updated but still failed to take virtualization into account. Shortly after, VMware announced its participation in the PCI council to help address virtualization within the PCI data security standards. While this is certainly good news and will help tighten up the security standards around electronic credit card payments, the outcome of this announcement remains to be seen. The following are a few improvements that shouldn’t be too difficult to implement right away:
1) First and foremost, the PCI council needs to recognize virtual hosts and include them in the scope of the standard if any of the virtual machines (VMs) that reside on the virtual hosts fall within the boundaries of the standard. Currently, any server, network or device that has anything to do with cardholder data would be included in the standard and any audits that occur.
Additionally, if any virtual machine is included in the scope, then all of the virtual machines on a host should be considered in the purview of the standard because they all reside on the same physical server. Finally, as virtualization allows for VMs to be easily moved between host servers for failure recovery and load balancing, all of the virtual hosts in a cluster should be included within the boundaries of the standard as well.
2) Clarify the confusing item (2.2.1) that dictates that you can only implement one primary function per server. All they have to do is exclude virtual hosts from this item.
3) Most of the security items that are listed in the standard can be applied to virtual hosts as well. This includes things like audit logging, password policies and applying vendor patches.
4) Address virtual networking. Ensure that the security settings on virtual switches do not allow things like promiscuous code, forged transmits and MAC address spoofing.
By simply addressing these four areas, the Payment Card Industry (PCI) standard would be moving in a better direction. From there the council could delve deeper and address other specific areas on virtual hosts using some of the existing security guidelines. Another distinction it should make is between bare-metal and hosted virtualization products. Hosted virtualization products are typically less secure because the underlying operating system is not optimized for virtualization. As a result, they should be subject to tighter scrutiny and control.
Virtualization administrators are in a unique situation where older operating systems (OSes) can potentially “live forever” in the virtual world. While we may not wish to enable older OSes to remain in our environments indefinitely as virtual machines (VMs), situations arise where we need to do just that.
Recently, I had a situation where an older OS had been removed from the installable toolkit platform — in this case it was VMware Tools. The older operating system, Windows ’98, had been removed from the VMware Tools installation with the release of VMware Server 2.0. While the need for a Windows ’98 virtual machine is rare, it does exist.
To solve the immediate problem, I was able to install a VMware Tools .ISO image from the 1.0.3 version of VMware Server that I’d been using on another host system. Once installed, the older tools are listed as ‘out of date’ as expected, but the basic features of driver optimization are present on the guest VM.
At first this dilemma did not appear to be much of an ordeal, but it started an important thought process. While Windows ’98 was the first occurrence of platform removal from a guest toolkit installation that I have observed directly, I don’t expect Windows NT or 2000 guestOSes to be that far from the chopping block of supported platforms.
One way to prevent this issue is to hold onto the tools installations for each platform of the hypervisor. VMware Tools, XenTools and Hyper-V Integration Services all exist as virtual CD-ROM .ISO images that you can hold onto for re-installation on another guest VM. Also, keep in mind that there may not be support from the host side either, so check to see which supported guest operating systems are available.
As you might expect, configuring an environment in this fashion may be met with some skepticism, as it could possibly divert resources. As a result, it may be worth placing this type of guest workload on a free hypervisor like VMware Server or on a similar lower tier of virtualization and storage. Having a flat file backup (.VMDK or .VHD) of the VM is a good idea as well.
While this situation is less than ideal for truly obsolete guest operating systems, the rare instance may arise where archiving toolkits can prove very beneficial.
I was just a teenager visiting family in France when I saw my first Minitel, France Telecom’s widely distributed teletext terminal for looking up phone numbers, viewing train schedules, and perusing naughty (!) message boards. While it looks hopelessly archaic now, in those pre-Web days, it was très cool.
Now I hear France Telecom is at it again through its subsidiary Orange Business Services. But this time, instead of targeting every French man and woman, it’s targeting small and medium-sized business (SMB) users with hosted IT services based on virtual desktop infrastructure (VDI) and low-cost access terminals (i.e., thin clients).
Judging from its website, the OBS Forfait Informatique seems to be based on Citrix XenDesktop, and starts at 99€ (about $125) per user, per month for a basic Microsoft Office pack. Virtual desktops can be accessed from existing desktops, or if you’d rather, OBS will subsidize a thin client from Wyse Technology, much in the same way cell phone carriers will give you a phone when you enter in to a long-term contract. Tarken Maner, Wyse CEO, tells me that Australian carrier Telstra is engaged in a similar project with Google to offer IT services to SMBs.
The idea that cable, telephone etc. providers might someday start offering hosted desktop services isn’t exactly novel — it’s certainly a logical progression — but is nevertheless an interesting development. How long can it be before the France Telecoms, Verizons and Comcasts of the world set their sights back on regular consumers, and offer virtual desktops as a monthly subscription, along with phone, cable and internet? Now that would be très très cool.
I received an email the other day from Wayne, Pa.-based SunGard Availability Services outlining some “essential” steps for addressing virtualization security challenges. In their email, the company urges users take certain measures, including installing security software, to make sure their virtual machines (VM) are safe from security threats.
There are many virtualization security products on the market today, yet reports of major VM security breaches are nil. In fact, the largest virtualization vendor, VMware Inc., asserts that its software is completely secure – possibly more secure than physical machines.
And even though the majority of VM security breaches I’ve heard about were hypothetical, performed by scientists through demonstrations or at hacker conventions, not in real data centers, I still receive a steady flow of press releases and product announcements addressing VM security issues.
So now, when I see security vendors warning users about un-named threats they need to prepare for, I am reminded of the U.S. Homeland Security Threat Level warning system.
Unfortunately, there are no published criteria for the threat levels of the Homeland Security system, so there is no way to tell whether the current threat level is accurate. And by the way, the threat levels have never been green or blue.
Because of this, the system can be manipulated by government officials. For example, during the Presidential election of 2004 when Republican President George W. Bush was running against Senator John Kerry, the Homeland Security Threat Level was bumped up, prompting some academics to speculate this was done by the Bush administration to scare voters into re-electing him. If so (and we will never know), it worked.
Unfortunately, decisions based on fear are usually not well thought out.
But I haven’t heard of any 9-11-style attacks on virtual infrastructures, and the virtualization users I speak with aren’t convinced they have anything to worry about. The thing that gets people to buy into virtualization security software is that haunting “what if” question that makes everyone default to the”better safe than sorry” mantra. After all, there is no harm in taking proactive steps to protect against the unknowns – just in case.
For instance, according to this article on the security benefits and risks of virtualization, “the [virtualization] drawback is based on fear of threats that aren’t around today but could become serious problems in the future.” Natalie Lambert, a security analyst with Cambridge, Mass.-based Forrester Research, continues in the article:
“One big concern is about what could happen if a flaw were found in a hypervisor, which would give attackers access to thousands of desktops sitting on a virtual server…That’s not a reality today, but it’s certainly a fear for the future.”
And as Sunguard said in its email, “With many organizations focusing on virtualization benefits, they must also examine core risks before it is too late – meaning security needs to be built in from the start.”
It is why we buy life insurance and car insurance and fire insurance for our homes. (Those damn what ifs and their expensive safeguards).
So, for the paranoid among us, check out SunGard’s suggestions for securing your virtual infrastructure here. As they say, better safe than sorry, right?
Integration Services is Microsoft Hyper-V’s installation interface on guest virtual machines that is designed to optimize the drivers of the virtual environments and provide the best experience. Here is a rundown of what you want to know about Integration Services when getting started with Hyper-V.
- Integration Services are installed via virtual CD – For default installations, the C:\windows\system32\ path of the Hyper-V server contains the guest.iso file. This virtual CD provides the installation of Integration Services and is launched from the Action menu on the virtual machine as shown below:
- Integration Services are native on some platforms – Selected releases of Windows Vista and Windows Server 2008 are Integration Services aware and do not need to be installed specifically.
- New Services and the Control Panel – Integrations Services shows up as Hyper-V Guest Components in the Control Panel and installs five Windows-based services. These are Hyper-V Data Exchange Service, Hyper-V Guest Shutdown Service, Hyper-V Heartbeat Service, Hyper-V Time Synchronization Service and Hyper-V Volume Shadow Copy Requestor. These Windows-based services exist in the task manager as vmicsvc.exe.
- Ease of Use – The installation of Integration Services permits the full use of Hyper-V Manager through remote desktop connections. Without this installation, interaction with the guest VM within Hyper-V Manager through a remote desktop from a different system will not permit mouse use.
Installing the hypervisor driver packages, such as VMware Tools, VirtualBox Guest Additions or Hyper-V Integration Services is always a wise decision in order to optimize the experience for guest systems. The configuration and setup of Integrations Services is very light, and can be managed much like the other driver packages. More information on Hyper-V can be found on the Microsoft website.
Today, as I read coverage of the VMware Mobile Virtualization Platform, or MVP, on the virtualization.info blog, I noticed a confusing tidbit. In his post, Alessandro Perilli states, “The fact that customers can run a Windows XP virtual machine [VM] on their phones doesn’t mean that it’s usable.”
While I totally agree with Perilli’s skepticism about the usability of Windows XP on a phone, I have to quibble with one point. In fact, VMware MVP won’t enable customers to run Windows XP on their phones at all.
Correct me if I’m wrong, but as a hypervisor for the ARM processor, MVP can’t run a Windows XP virtual machine, because XP is designed to run on an x86 processor, even if it is virtualized. (It can, however, run a Windows CE VM.)
If MVP presented an x86 emulation layer, it would be a different story. But that’s not the story VMware told me.
The more I think about it, the more I like Citrix’s proposed ICA client for the iPhone. While the Citrix approach doesn’t do much for embedded device manufacturers, it gets at end users’ goal of accessing their desktops and all the data that resides on them in an elegant, secure fashion. And if the effusive comments on the Citrix blog are any indication, I’m not alone in thinking that way.
Last week I noticed that the Payment Card Industry’s Data Security Standard (PCI-DSS) was recently updated on October 1, 2008, from version 1.1 to 1.2. PCI-DSS is a security standard set forth by a conglomerate of all the major credit card companies and is designed to protect cardholder data. As a result, any company that accepts credit cards is forced to comply with it.
About six months ago I wrote that the PCI-DSS standard did not specifically address virtual environments, and instead only focused on servers and networks that are directly involved with cardholder data. In other words, the specification dictates what must be done to secure a server that may store or process cardholder data, but if that server happened to be a virtual guest the host server would not be considered in the scope of the specification. Subsequently you could secure a virtual guest all you want, but if you do not properly secure the host server you could easily compromise the virtual guest regardless of how it was secured.
I downloaded the summary of changes document that specified all of the changes that were made from version 1.1 and 1.2, anxious to see if they had finally added parameters for virtual host servers. Out of the 14 pages of changes, there was still no mention of virtualization technologies in the specification. Surprised by this, I searched through the whole version 1.2, 72-page specification document for the word virtual and found only one instance of it for virtual private network.
I am puzzled as to why they would continue to ignore virtualization. After all, isn’t just about every company virtualizing in some fashion these days? Are the people that write the specification parameters just ignorant of what virtualization is, and that it has a direct impact on their regulations? Or are they just trusting that we are all securing our virtual hosts properly and there is no need to address them? If that’s the case then they have misplaced a critical amount of trust as I am sure there are a great many virtual environments that are not properly secured. Likewise, ignoring virtualization completely greatly reduces the effectiveness of their efforts to secure environments that deal with cardholder data. It’s essentially fortifying everything within a castle, but leaving the front gate open.
It wouldn’t require a great deal of effort for them to address virtual hosts. A number of security specifications for virtual hosts already exist, such as cisecurity.org’s for VMware’s ESX. Let’s hope that they wise up and address virtualization in their next update of the specification. Until then their efforts to protect cardholders are not complete. I just hope that my credit card data is not lying on a virtual machine somewhere that resides on an insecure host server that is ripe for the picking. After all, why try and hack a single virtual machine when you can instead hack into a whole host and gain access to all the VMs and their data?