Brian Madden has a nice
overview
of Citrix' HDX family of technologies. HDX stands for "high definition
experience," and is part of the XenDesktop 3 release, set for sometime later
this month.
Brian points out that there really isn't anything new in HDX,
but believes it's good marketing for Citrix; nothing wrong with trumpeting what
you've got. As for XenDesktop, our
review
of the last version was glowing; I can't wait to see how Citrix improves the
latest iteration.
One of the best enhancements, from my point of view, is
that Citrix claims it can double your
VM
density.That means you can put twice as many virtual desktops on a server as
before, with (allegedly) no performance degradation. Anything that allows you to
improve VM density is going to be a hit, since it means using existing hardware
rather than buying new stuff.
Price for XenDesktop 3 is $75 per
concurrent user. The precise release date wasn't given.
(Personal note:
my blogging activity has been significantly impacted by a recent back injury I
suffered. I'll be picking up the pace shortly.)
Posted by Keith Ward on 02/13/2009 at 12:48 PM0 comments
They're baaack...the
rumors of
Cisco possibly buying VMware and parent company EMC. Virtualization.info has a
good
breakdown
of some signs that point to a possible acquisition.
It's all
speculation, of course, but there's solid fact-reporting behind the theory.
VMware had a very good Q4 in 2008, one of the few companies in the IT industry
to do so. Cisco is primarily a hardware company, but is intensely interested in
virtualization, as the upcoming
Nexus
1000V demonstrates.
EMC and VMware seem to be getting along better
since Paul Maritz replaced Diane Greene last year, but that doesn't mean that
the storage and virtualization giants are a good fit. Still, given VMware's
value, EMC might want to hold on to it -- unless it's been waiting for the right
time to sell VMware for some time, and believes that VMware's value is at its
zenith.
Stay tuned.
Posted by Keith Ward on 02/12/2009 at 12:48 PM2 comments
Virtualization security is in somewhat the same boat as
virtual desktop infrastructure: lots of discussion, but less actual
implementation. It's my belief (unverified and mostly anecdotal) that admins
think their standard security measures will protect their virtual environments.
Or maybe they do, in the dark recesses of their minds, know that they need to
beef up their virtual security, but lack the time or guidance necessary.
I can't help with the time part, but in terms of guidance,
Microsoft is doing its part with the recent release of a
Hyper-V
security guide. Note that you need a LiveID and have to join the beta
program. The Windows virtualization team blog
reports
that three main areas are stressed:
- Hardening the Hyper-V server role.
Of course, one
reason for a server core implementation is increased security, but it's good to
go into more detail on installation and configuration
parameters.
- VM management and delegation. Unless you're
a one-stop shop, you'll need to know the safest ways to delegate management
functions to others, especially in the areas of administrative
access.
- Protecting the VMs themselves. That includes auditing information, encryption and file
system permissions.
If you are testing Hyper-V or using it in production,
I'd call this a must-do. And soon.
Posted by Keith Ward on 02/11/2009 at 12:48 PM0 comments
Have you ever
wondered which hypervisor is fastest? I'm talking pure, raw speed here. No
management tools, no third-party utilities, nothing that gets in the way of a
sprint to determine which hypervisor can do the most in the least amount of
time.
Well, I've wondered. And decided to do something about it. Rick
Vanover, our new online columnist, put the base hypervisors from the three
leading vendors to the test for the cover story of
Virtualization
Review magazine's Feb./March isse. We put VMware's ESX, Microsoft's
Hyper-V and Citrix' XenServer through the ringer, to see how each handles
varying workloads. The results were not at all what I expected -- and I think
they'll surprise many of you, too.
Also in the same issue are a couple
of features I wrote -- one is a case study of a virtual desktop infrastructure
(VDI) implementation at the University of Maryland. The administrator of one of
the school's colleges decided he was tired of relying on Sneakernet, and started
examining VDI solutions. He eventually settled on one, and has been using it in
production for some time now. They use some extremely resource-instensive,
high-level programs there, the kind of applications that can reduce a laptop to
a smoldering heap of rubble. How those apps work in a real-world VDI scenario is
eye-opening.
The other feature is about processor-based virtualization.
As you may know, Intel and AMD are in a dogfight for virtualization market
share, and are continually upgrading their chips to allow for speedier, more
efficient virtualization. The story discusses the lay of the land, and what we
can expect to see in the future.
The next issue also unveils a major
change to our "Upfront" section. Instead of news and analysis, I've decided it
could be better used for product reviews. So many of these products are
unfamiliar in the market, and many of you haven't had a chance to use them yet.
Beefing up that section with more product coverage makes it more useful to you,
I believe. If you disagree,
let me
know. I take reader feedback very seriously.
We also have our usual
great columns from Danielle Ruest and Nelson Ruest (our Virtual Architects), and
Chris Wolf. It's going to be a great issue, one that will get folks talking. As
an editor, that's what you always hope for.
If you have any ideas for
future stories, or feedback about what you love and hate about the magazine, I
want to hear from you!
Posted by Keith Ward on 02/10/2009 at 12:48 PM1 comments
A couple of weeks ago, I
blogged about a survey on virtualization.info that shockingly showed more folks using Microsoft's Hyper-V than VMware's ESX.
The thrust of my comments was that we needed a lot more information about the survey to be able to determine the legitimacy of the results. Well, Alessandro Pirilli, the owner/operator of the Website, has provided that information. What he revealed tells me that his results aren't to be trusted.
There was no filtering of survey results, and anonymity was granted. Alessandro says himself that anyone could vote as often as they wanted, including vendors. Really, you don't have to go any further than that to throw out the findings.
Alessandro, to his great credit, also says that people from the Microsoft and VMware domains both voted heavily in the survey -- but much more heavily for Microsoft than VMware. He gives the exact totals, too.
Despite the fact that Microsoft cheated a lot more, both companies tried to stack the deck -- and for that, both vendors should be ashamed. Thanks, guys -- instead of a survey that could have given us valuable insight, your transparent attempts at ballot-box stuffing invalidated the results.
Naturally, the validity of the responses to the other poll questions has been destroyed as well. It's a shame, because it's obvious that Alessandro put a lot of work into this effort. I'm sure he learned a hard lesson from this on how to properly do a survey. I also hope that the cheaters at Microsoft and VMware learned a lesson -- but I'm much less confident about that even than the survey findings.
Posted by Keith Ward on 02/03/2009 at 12:48 PM1 comments
Blogging, in a way, is the lifeblood of the virtualization
community when it comes to information sharing (other than
Virtualization
Review in print, of course). More so than any other IT industry, blogging
is the main way readers get informed about what's going on. I think it's because
it's such a new field that the old methods just didn't take root as quickly.
Blogging is one of our key focuses here on the Website, too.
And starting this week, we've got some very cool additions to our blogging
efforts. We've got three new bloggers coming on board, which is going to upgrade
our content substantially.
The first two bloggers write as one team. I'd
like to welcome Nelson Ruest and Danielle Ruest, our "Virtual Architect"
columnists, to the
Virtualization Review blogosphere. If you follow
their column and stories in print and online, you already know that they're two
of the best in the business. Now they bring their insights on all manner of
virtualization topics to our Website twice per week.
Our other new
blogger is Rick Vanover, a prominent writer on virtualization topics for
numerous Websites. Rick's new blog will be called "Everyday Virtualization,"
because that's what it's about; how to use virtualization every day in your job.
Rick has written a number of articles for the magazine in the past, and I'm
delighted to add his real-world, in-the-trenches take on
virtualization to our lineup. Like Nelson and Danielle, he'll be blogging
twice weekly.
In addition, Rick will be writing a monthly, online-only
"how-to" column focusing on getting things done virtually. Rick works with this
stuff everyday, and he'll be showing you, step-by-step, how to get the most out
of your virtualization environment. You won't be able to find it anywhere else
on the Web. I'll update you each time a new column comes out. One related note:
Rick wrote the upcoming cover story for our February/March print edition,
comparing raw hypervisor performance in a way you haven't seen before. It's
going to be a must-read in the virtualization community.
Please join me
in welcoming Dannielle, Nelson and Rick to our blogging family. It's going to be
exciting!
Posted by Keith Ward on 02/02/2009 at 12:48 PM0 comments
A new virtualization company opens its doors
today, and is offering something I haven't seen before in the virtualization
world: memory virtualization.
Wait a minute! you shout. Memory
has been virtualized for a long time. VMware's ESX does it; so does Microsoft's
Hyper-V, Citrix' XenServer, Virtual Iron and everyone else. They all
pool server RAM and make it available to virtual machines (VMs) on an
as-needed basis.
That's what I told
RNA networks Marketing VP Andy Mallinger
in a briefing last week. But RNA is doing something different: specifically,
its Memory Virtualization Platform (MVP) and first product,
RNAmessenger, isn't designed to work in a virtualized
environment. It's for very high-end servers that normally wouldn't
be virtualized due to their workloads.
RNAmessenger works in
basically the same way as the other guys. It pools memory, but in this case
makes it available only for physical servers that need it. The company is
claiming perforance improvements of 500 percent and even more in initial use
cases with financial companies that rely heavily on electronic trading. The
first company using it is a gigantic hedge fund that RNA would not identify by
name. Mallinger did imply, however, that the company is beyond happy with the
product.
A performance improvement of 500 percent plus? Those are heady
numbers, to be sure, but RNA has promised to provide me with customers who will
verify the figures. Be assured that I will follow up, as those numbers
are amazing if true. It claims to do this through reducing latency and
increasing throughput.
Right now, RNAmessenger is for use on
various flavors of Linux, including Red Hat, Fedora and SUSE, and is
x86-based. It requires the MVP hardware appliance and
the RNA messenger software.
As a company, RNA looks as if
it's started off on the right foot. It's received
$7
million in Series A funding, a significant amount of
start-up capital in this economic environment. It's got a strong leadership
team, and a product that fills a niche I hadn't even considered before, but
makes perfect sense.
RNAmessenger is available starting today. If
you become a customer, please
let me
know about your experiences.
Posted by Keith Ward on 02/02/2009 at 12:48 PM2 comments
So, Microsoft has finally done it -- announced
layoffs. I can't ever remember Redmond actually laying off employees for economic reasons (although I'm more than willing to be corrected on this point.) It just goes to show how serious this recession is shaping up to be.
It's not like it's a huge percentage of the workforce; 5,000 total out of 90,000-plus is less than sweeping. But it is an indicator of two things: a) That Microsoft's sales are plunging, and b) That Microsoft expects this trend to continue into at least the immediate future -- meaning, probably, the rest of this year and maybe beyond.
How will this affect its virtualization efforts? My guess is not at all. Microsoft knows that virtualization may be one of its best bets in the recession, since it's a money-saving technology. Note also that the server and tools business showed healthy growth, in contrast to nearly every other division. I take that as meaning, at least in part, that Windows Server 2008 is doing well in the market. And of course, Hyper-V is a key feature -- maybe the key feature -- of Windows 2008.
That's why we're seeing such rapid progress on Windows 2008 R2, already out in beta. It includes, naturally, the R2 beta of Hyper-V; Hyper-V Server is on the same path. I wouldn't expect any of Microsoft's virtualization initiatives, including Virtual Machine Manager and its various desktop virtualization plays, to be affected one iota because of today's news. In fact, if you're a Microsoft developer, it might be worth your while to get attached to a virtualization project.
Posted by Keith Ward on 01/22/2009 at 12:48 PM3 comments
Well, this should shake up folks in the virtualization industry. Alessandro Pirilli at virtualization.info has posted the results of a "
Virtualization Industry Survey" he conducted over a month's time. He says he received 1,050 responses, and those responses are interesting, to say the least.
The finding that made me sit up, and I'm sure everyone else, was Question 6, "What hardware virtualization platform do you implement?" It was a multiple-choice question with eight hypervisor choices, including the Big 3 of Microsoft's Hyper-V, VMware's ESX and Citrix' XenServer.
To the shock of, I'm sure, everyone who read it, Hyper-V had a commanding lead, with nearly 600 people saying they're using it. ESX was way back in second place, with around 475 votes. XenServer was a blip, with fewer than 100 votes, and the others barely registered. Note that the bar graphs Alessandro uses are imprecise, making it hard to estimate with great accuracy the tallies; but they're close enough.
Now, anyone who follows this industry will be immediately suspicious of these results. ESX as a platform has been around for years, and Hyper-V less than a single year. In addition, VMware is the clear leader in this industry in number of installations; on that, I would stake my reputation and job. It just isn't close.
Comments posted with the survey results are already questioning their validity -- and I am among them. It's completely counter-intuitive, raising red flags all over my simple mind. I don't know how scientific these results are, and Alessandro's questions aren't designed, from what I can tell, to dig down into the reasons for the numbers.
For instance, there was no information listed on commercial implementations vs. private ones. Also, Hyper-V is part of Windows Server 2008; are those respondents using Hyper-V, or did they respond in the affirmative because they've installed Windows 2008 on their networks and Hyper-V is part of the package, even if they're not using it?
Other questions: How did Alessandro filter out multiple votes? What was the voting process? How were vendor votes weeded out? What is the margin of error on his results? When I was editor of Redmond magazine, we did an annual salary survey (which is still being done.) We hired a private survey company to do the poll; we wanted professionals to handle the data, knowing we'd get more accurate results.
Note that I am not saying that Alessandro's results are meaningless; not at all. I am saying we need more information to know how to judge the findings. I hope he opens up about his methodology.
Posted by Keith Ward on 01/20/2009 at 12:48 PM4 comments
Microsoft is pushing a little further into the desktop virtualization space with the
release of the first beta of Microsoft Enterprise Desktop Virtualization (
MED-V).
Quite frankly, I'm not sure what to make of this.
First, it's part of the Microsoft Desktop Optimization Pack (MDOP).
Why Microsoft keeps releasing virtualization technologies (like its application
virtualization tool, App-V) in this restricted way doesn't make sense to me. What if you want MED-V or App-V, but don't need or want Software Assurance? Sorry, you're out of luck, says Microsoft. Either you become part of our revenue stream, or you can't have. This really, really rankles me.
Is Microsoft within its rights to do this? Sure. It's totally stupid, in my opinion, but Redmond should be free to run its business however it wants. As consumers, of course, we're equally free
to say "Bite me, Microsoft," and get VMware's ThinApp or Citrix' XenApp without having to enter into a contract.
The other thing is that it only works with Virtual PC. Remember that it's a hosted hypervisor, meaning it exists as an application in the base OS. Virtual PC, although a nice product for what it
does, is limited in functionality, and by its nature as a hosted hypervisor, slow. Also note that Windows 7 isn't supported in the first beta. Maybe it will be in the final version, expected out sometime in the first half of this year.
All in all, this release gets more of a yawn from me than anything else. Lots of exciting stuff is happening in the world of desktop virtualization; too bad this didn't add to that excitement.
Posted by Keith Ward on 01/16/2009 at 12:48 PM8 comments
If there was any doubt before now, there should be none after today: Microsoft wants to do more than compete in the virtualization arena -- it wants to win.
If you hadn't noticed (and apparently many folks didn't -- it's been almost like a news blackout on this announcement), Windows Server 2008 R2 is available in a public beta. The feature that matters most to readers of this blog is that it includes, for the first time, Live Migration as part of the upgraded Hyper-V.
Of course, how good a product Live Migration is at this stage is another matter entirely. Hyper-V has been well regarded by almost everyone who's used it. That's not to say that it's up to the level of ESX or XenServer in terms of performance or functionality; but for what it's intended to do, which is -- in Microsoft's own words -- primarily server consolidation, it works quite well. If Live Migration lives up to the standard of the initial Hyper-V release, we should expect that it will be a very solid v1 product.
But even the fact that it's out this early shows how badly Microsoft wants to horn in on VMware's turf. Live Migration wasn't supposed to be live for another year or so; with the beta out so early, we're likely to see a final product by mid-year. What that tells me is that Microsoft knew it was losing out to VMware every day that it didn't have a live migration-enabled hypervisor.
I'll also watch with amusement as Microsoft, which has spent a couple of years telling everyone that Vmotion (VMware's implementation of live migration) wasn't all that big a deal, will suddenly realize the huge importance of live migration in a datacenter setting.
Another nice improvement in Hyper-V R2 is that it now supports up to double the number of logical processors, from 16 to 32. Scalability has just taken a big jump. Also, note that Hyper-V Server, the non-Windows 2008 version, has the same upgrades.
It all adds up to a nice set of improvements. Will it be enough to erode VMware's market share? We'll see.
Posted by Keith Ward on 01/09/2009 at 12:48 PM4 comments
VMware has a new Chief Operating Officer. The company
announced yesterday that Tod Nielson, a software industry veteran with more than 20 years experience, becomes the new COO.
What happened to the old COO? Nothing -- it's a newly-created position. It seems a little odd to me that a company the size of VMware didn't have a COO before this.
CEO Paul Maritz said in the release that Nielson will focus on the business aspects of the company, freeing him up for more product strategy and development work. Maritz appears to be a geek at heart (not a bad thing), so this is probably a load off his back.
The best part of the press release, however, was the last line:
"The company's share price has ranged between $17.25 and $85.84 over the past 52 weeks."
Now that is what I call a roller-coaster ride!
Posted by Keith Ward on 01/07/2009 at 12:48 PM3 comments