This week, more than 23,000 attendees descended on San Francisco's Moscone Center for the 12th annual VMworld, the virtualization industry's biggest party. Here are some notes and observations from day one:
- First, the number of attendees is a new record. The day one keynote mentioned that the very first VMworld, held in San Diego, had about 1,200 attendees. So for those who were wondering if interest in virtualization is on the wane, those kinds of numbers should put those fears to rest.
- Related to that is the number of journalists covering the event. At an early press conference on Monday, the main press room was overflowing with journalists and analysts. I'm sure there were more than 150 in the room, and not all the journalists that cover the show attended the event. This is my fifth VMworld, and I don't remember this large a contingent of reporters. (Another related note: The amount of space set aside for journalists and analysts at the keynote was far short of what was necessary.)
- At the press conference, I asked a question about vCloud Air, VMware's hybrid cloud platform. It was about the Amazon Web Services (AWS) and Microsoft Azure dominance of public cloud, and what VMware needs to do to make its offering a legitimate player in this space. VMware's Bill Fathers, EVP and GM of Cloud Services, said he "disagreed with the premise" of my question, then went on to discuss how good the vCloud Air technology is and how it's growing.
When Fathers was done answering, the journalist next to me whispered that Fathers didn't answer my question. I talked to a prominent analyst afterward and he said the same thing. The reality is that no matter how good vCloud Air is, it's still in the "other" category for public cloud, well behind AWS and Microsoft.
- The day one keynote didn't feature CEO Pat Gelsinger. I'm used to a company CEO doing the first keynote -- it's typically the most well-attended one, and also where the company discusses its broad vision. (This may be typical for Gelsinger; if it is, please let me know. Gelsinger is giving the closing keynote address Tuesday morning.
- There were many announcements made at the keynote, but what struck me most was that there was almost nothing new unveiled. Instead, the company touted lots of upgrades to its cloud infrastructure products, its vision of the software-defined datacenter (SDDC), better automation and management capabilities and new opportunities for development, among other announcements. Overall, though, there was no "wow" moment.
That isn't to say that that the announcements aren't important; they are. But in terms of announcements that knock you out, it just wasn't there. VMware is moving ahead in important ways, and it seems clear that it's still a vibrant, innovative company that isn't afraid of change (that was obvious in its complete and total embrace of container technology, which it spent a lot of time on in the keynote presentation). But the applause from the audience during the keynote was mostly polite and muted for good reason.
Posted by Keith Ward on 08/31/2015 at 1:04 PM0 comments
Two weeks from today I'll be a VMworld, like many of you. I'll be covering news, talking to vendors, watching demos, standing in line; in other words, everything that makes up a typical show.
Attendees will get the full-court press from VMware, pushing its vision of "Ready for Any," but during my time there, I also hope to get insight on some issues VMware won't be promoting as strongly. Things such as:
- Sales figures for EVO:RAIL. There have been hints that EVO:RAIL, VMware's hyperconvergence solution, hasn't done well in the sales channel. VMware hasn't broken out details (anywhere that I've seen, at least) for sales, and has been pretty quiet on that front since launch. Concerns are that it's priced too high, and not as advanced as some of the competition. If sales continue to stagnate (assuming they are -- VMware may have a different opinion on what constitutes stagnation), will the company consider pulling it?
- Public and hybrid implementations of vCloud Air. VMware's cloud platform is well-regarded, and by all accounts has done well in the private cloud space. When it comes time to extend that into the public cloud, however, vCloud Air doesn't appear to be making great headway. Right now, Amazon Web Services and Microsoft Azure are the two primary choices for enterprises, and vCloud Air is somewhere… lower on the list. Even in the private space, OpenStack has all the momentum right now (which is why VMware has pushed out its own OpenStack integration). But given the growing importance of the cloud to IT operations, it's imperative that VMware have a strong story in the market, along with receptive customers. Is vCloud Air up for big changes that will allow it to scale better, for example?
- VMware's future. The company is the subject of constant rumors about whom might buy it, and who it might buy. More than ever, parent company EMC feels like an anchor around VMware's neck; to me at least, VMware seems like a more innovative, forward-thinking company than EMC, which maintains more of a traditional mindset. If I'm VMware, why do I want to be part of EMC? EMC itself doesn't seem to know what it is (witness the "federation" structure Re/code reports on), so how would it know what to do with VMware?
I'll be spending time at VMworld looking for some of these answers. If you see me there, feel free to let me know what you think, or hit me up @VirtReviewKeith.
Posted by Keith Ward on 08/17/2015 at 12:37 PM0 comments
As I wrote about recently, hyperconvergence vendor Nutanix and VMware have been battling over testing of their products. That battle took an interesting turn last week, when one storage testing site took Nutanix to task over its hesitancy to let the site test its appliances.
Performance benchmarking and testing, as you may know, is a sensitive topic for many vendors. They can fear testing for a number of reasons: apples-to-oranges tests, pristine test environments that don't account for real-world scenarios, bad testing methodologies and so on. And they're right to be cautious; testing can be skewed to give the results a certain vendor wants; in addition, many reviewers aren't very sophisticated about testing, and don't understand how to do it right.
No Cluster for You!
But Nutanix didn't do itself any favors with its recent behavior toward StorageReview.com. In an article titled "Why We Don't Have a Nutanix NX-8150 Review," Brian Beeler discusses the problems he had getting that particular Nutanix appliance. Here's his summary of the events:
"Nutanix sent an NX-8150 cluster for review, our testing revealed performance issues, Nutanix made several software updates over six months to improve performance, in June they asked to update our cluster with newer Haswell-based systems, we agreed, Nutanix then backtracked and refused to send the promised replacements."
It's a long saga, and details how Nutanix became increasingly wary of publishing the results of any performance testing unless it had final approval of StorageReview's article. Nutanix also asked StorageReview not to use industry-standard methodologies like VMmark and Sysbench for testing. "Until we have a mutually agreed upon plan, we ask that you not undertake any performance testing of the Nutanix product, or publish results of prior performance testing," Nutanix said in an email, a snippet of which StorageReview posted.
Mistakes Were Made
Lukas Lundell, global director of Solutions and Performance Engineering at Nutanix, responded in StorageReview's forums (Lundell fired the main shots at VMware's attempts to test Nutanix vs. VMware products). To his great credit, he apologizes for Nutanix's failure to properly handle its interactions with StorageReview: "… it appears like we definitely started off on the wrong foot, and it's very clear we mismanaged this situation. We didn't treat them like a customer, and that was a big mistake."
That's absolutely the right way to do it. Take your lumps where you should; no one will feel more negative toward a company that admits mistakes; on the contrary, they tend to trust that company more, rather than less.
It's also important to point out that numerous Nutanix customers responded on the same forum thread, and they were uniformly happy with their Nutanix experience. Here's a typical example, from "nathanmace": "Our Nutanix experience has been excellent, rock solid stability, great performance, and some of the best support I've ever dealt with."
But StorageReview's experience with Nutanix, combined with Nutanix's vitriol toward VMware, makes me wonder why the company appears to be so worried about anyone publishing test results of its products (and Nutanix's attempts to control what's actually published), and so eager to quash any attempt to evaluate the performance of its products or compare it to others. It's especially puzzling in light of the excellent reputation Nutanix generally has in the industry; it's not like it's some fly-by-night startup with no track record of happy customers.
Questions That Need Answers
Bob Plankers, an analyst with The Virtualization Practice
, has some of the same questions I do, which he posted on his personal blog
: "Why can't I run a standard benchmark like VMmark on a Nutanix cluster? Why can't people share performance results? If I bought one of these would I be able to talk about my performance? Why is Nutanix uncomfortable with performance results?"
The reality is that nearly any test that anyone wants to run could be blasted as somehow invalid or biased, if one wants to dig into the testing details to the nth degree. Yes, there's a right way and a wrong way to test, and the wrong ways should be called out. But nitpicking results in good tests -- and good testers, like StorageReview -- because of slight variations here or there, seems unnecessarily defensive.
The Beginning -- Not the End
The last thing to say about performance testing is that test results are a starting point, not an end point, when evaluating storage, networks, infrastructure or anything else in IT. They should never be the final word for your
datacenter. They are one factor of many to consider. Nutanix has many fans. So does VMware. They have them for a reason.
Posted by Keith Ward on 08/10/2015 at 1:09 PM0 comments
So, EMC is spinning off VMware. Or not. Instead, it could fully integrate VMware. Or maybe it's the other way around -- perhaps VMware will buy out its parent company, and absorb EMC. Or not.
Welcome to the wild, wacky world of Elliott Management. Elliott is known for its hyper-aggressive strategy of getting a stake in a company, suggesting lots of changes to increase "shareholder value," then taking further -- often drastic -- steps to implement those changes.
And, according to Re/code's Arik Hesseldahl, the latest buzz is that EMC is considering the benefits of being purchased by its own subsidiary. The deliberations are part of a "partial response to pressure" from Elliott, Hesseldahl writes.
The idea of EMC selling off VMware, which it purchased in 2003, has floated around the virtualization industry for years. The businesses are very different, while also having some commonalities. EMC focuses on storage, while VMware's core concern is datacenter virtualization and management.
Both companies, however, have significantly broadened their technology portfolios over the years: in addition to VMware, EMC owns security giant RSA, and data analysis company Pivotal (run by former VMware CEO Paul Maritz). VMware is now heavily invested in cloud computing and mobile device management, as well as hyperconvergence with its EVO:RAIL appliances. Elliott often sees diversification as a negative, complaining that companies can lose the focus that made them successful in the first place.
A Downstream Merger?
The idea of VMware buying out EMC is much newer. That would be what's called a "downstream merger," Hesseldahl reports: "In this instance, VMware would issue new shares in exchange for EMC shares in combination with cash raised from the issuance of new debt."
If that occurred, any moves VMware would take with a company as large as EMC is pure speculation. VMware is pushing heavily into software-defined storage (SDS) as part of its overall strategy of the "software-defined datacenter," or SDDC, which attendees to VMworld later this month will undoubtedly hear much more about. EMC is more of a traditional storage company, so an integration of its products into new VMware offerings could be exceptionally complex and difficult.
Competition vs. Integration
Another problem is that the two companies now compete against one another in several areas, including cloud infrastructure. Can a company divided against itself stand? That's another critical issue to be resolved, whether or not one buys out the other. For the sake of simplicity, it would be better for EMC to simply sell off VMware, which is why that's been the most-predicted outcome over the years.
If an assimilation occurs (either way), another significant hurdle would be how to integrate the two different corporate cultures. VMware has always operated independent of EMC, for the most part; it's been a point of pride for VMware, in fact. I'd argue it was necessary, too, since VMware understands the virtualization market better than anyone, and having EMC interfere would lead to nothing but negative outcomes. Would both companies suffer if they were fully integrated? It's certainly a possibility.
Game of Thrones
At this point, no one knows how this will all shake out. Elliott is good at getting what it wants, and right now it appears to want the status quo to go away. Will VMware and EMC be able to resist Elliott's suggestions? Do they even want to? Or is Elliott's intrusion the justification the companies have been seeking to go in a completely different direction?
One thing is sure: It will be fascinating to watch this all play out. What's your take? Should the companies separate, or should one buy out the other? Let me know in the comments section.
Posted by Keith Ward on 08/07/2015 at 9:39 AM0 comments
VMware has released a whitepaper that shows performance increases in vSphere 6.0 over vSphere 5.5. It has some interesting results -- how real-world valuable they are is another matter.
One of the most important things to notice about this document is on the very first page: It's called "Technical Marketing Documentation." So it's been produced for the specific purpose of selling more vSphere 6.0. That doesn't necessarily mean that the results are skewed to show better performance, but it does mean you should approach these figures with a skeptical -- though not cynical -- eye. Just keep in mind that VMware has an agenda beyond just providing helpful performance information about the latest version of its flagship product.
The whitepaper looks at performance across vCenter Server, before tackling core platform, storage and network functions.
vSphere Web Client
It's interesting, and not at all accidental, that the very first performance-related enhancement the document mentions is the vSphere Web Client. VMware has been trying for some time to move users off the built-in C# client, but complaints about the Web Client remain persistent and consistent. If it truly has improved on the areas mentioned -- "the login and home pages, summary pages, networking pages, related objects lists, general navigation, performance charts, and action menus (right click)" -- many admins will be made very happy. Using a Web Client makes the most sense these days, but only if it's comparable to the built-in one.
vCenter Server 5.5 and 6.0 performance was compared using a Microsoft SQL Server 2012 database of various sizes: small, medium and large inventories. VMware reported sharp gains overall, with 63 percent improvement for a small inventory, and 73 percent and 67 percent gains for medium and large inventories, respectively. For a medium SQL inventory, the throughput from vSphere 5.5 to 6.0 jumped from 471 operations per minute (OPM) to 1,292 OPM. At the same time, operational latency declined, from 60 percent for a small inventory to 84 percent for a large one.
Another critical area for VMware admins looking to move to vSphere 6.0 concerns storage performance for the upgrade of its Virtual SAN (VSAN). In this area, VMware's testing looks like a win as well. "Among many significant improvements, performance has been dramatically improved for both hybrid and newer all-flash configurations," the report states.
For VSAN Hybrid, the numbers show a range of improvements from double the performance of IOPS per host over VSAN 5.5, to 3½ times the improvement for IOPS on a 64-node cluster. Latency has also improved (i.e. decreased) by 40 percent, according to the testing. The results for all-flash VSAN arrays are mostly incomplete (lots of "Pending" numbers here), but one should expect those numbers to be even better.
Networking improvements are similarly impressive, although they're highly dependent on the configuration; VMware's tests used Windows Server 2012 and Windows 8 as guest operating systems.
The thing to remember about all these figures is they're performed in an environment that's different from yours; in most cases, an enterprise won't see these types of gains. Still, it's instructive to see that upgrading to vSphere 6.0 will likely result in significantly better performance for your shop. Have you put vSphere 6.0 in place yet, running in a live production environment? I'd love to hear your feedback, at either [email protected], or @VirtReviewKeith.
Posted by Keith Ward on 08/05/2015 at 10:05 AM0 comments
More In This Series
It's been a fun trip down virtual memory lane, as I've strolled through the pages of Virtualization Review's very first issue, from April 2008. As I look back, it's been fascinating to see where our predictions about the future were right and where they were wrong, and how very much things have changed in the seven-plus years since we began publishing.
I'm not sure that anyone at that point saw the rise of software-defined this, that and the other. Things like storage virtualization had already been birthed, so even though it's changed a lot, it's not like it was completely unexpected. But software-defined networking; well, that was something that turned me upside down like a theme park ride.
I admit also that I didn't see the rise of hyperconverged appliances. I assumed that datacenters would, for the most part, remain the way they were back in the days when I worked in a world-class datacenter at E.W. Scripps: distinct silos for everything, including servers, networks and storage. It's a testament to the never-ending innovation in this space that such things have occurred in a (relatively) short amount of time.
In terms of vendors, they haven't changed all that much: VMware still leads the pack, while Microsoft and Citrix are still major presences (although they've reversed positions from what I expected in 2008: I thought Microsoft wouldn't make much headway with Hyper-V, and that XenServer would take off). HP, Dell, Oracle, Red Hat, SUSE and the like are still doing cool things and staying highly relevant. Contraction has occurred, especially in the storage space, which was to be expected.
KVM has done really well, better than I expected way back when. Xen has found its niche, but not gone much beyond it, while ESXi has continued its dominance and Hyper-V has done extremely well. As for hypervisors, that's about it. Though there are others out there, they're mostly white noise.
In terms of disappointments, I'll say that virtual desktop infrastructure (VDI) hasn't had anywhere near the impact I thought it would by now. It certainly is important in many settings, but hasn't become ubiquitous like numerous vendors predicted. This doesn't fall under the heading of breaking news, but it's still instructive -- remember to keep your perspective when you're in the middle of the hype machine.
Finally, the big thing we missed in issue No.1 is cloud computing. It certainly was known in 2008, although at the time I wondered if it would ever go mainstream. As everyone reading this knows, it's done that in a big way. In fact, it's switched places with VDI: many thought VDI would be the big deal by 2015, and cloud computing would be a solid niche technology, but no more. I was one of those (boy, has this been a humbling article to write).
Now, of course, cloud computing is changing the industry -- in much the same way, in fact, that virtualization did, lo these many years ago. I don't see its momentum slowing any time soon, although it will eventually reach that point. In the meantime, every company has to figure out a strategy for using it, and there are tons of vendors out there to help them with those decisions.
Looking back, I'd call issue No. 1 a big success. Not because we were right about everything; far from it. Its success lies in the fact that it marked the entry of something brand new: the only print publication covering virtualization as an industry. There were Web sites, yes; plenty of those (as there still are, of course). But it was the era of contraction in the magazine business, and IT publications were especially at risk, because so much of the audience was online. In fact, many of the competitors in my company's specialized field of business-to-business IT publications have gone out of print, and it's been a sad thing to see. But Virtualization Review has survived as both an online and print publication (although it's not published as regularly as it used to be).
Being first at something is way cool, which is why Issue No. 1 stands out in my mind so clearly. How often do you get to be part of a pioneering effort? Well, I was when issue No. 1 rolled off the presses. I was honored to be a part of it then, and I'm still honored to be a part of it now.
Posted by Keith Ward on 07/20/2015 at 9:52 AM0 comments
More In This Series
Continuing my analysis of the changes between Virtualization Review's first-ever issue from April 2008 and now, we come next to the topic of storage virtualization. This field has changed as much as any since that edition of the magazine hit newsstands.
The article, written by Drue Reeves (then an analyst with Burton Group, now chief of research at Gartner, which bought Burton Group in 2010), was called "The New Virtualization Frontier." Drue started off with a grabby lead: "If 2007 was the year of server virtualization, then 2008 may be the year storage virtualization finds its way into many data centers." He was correct that if server virtualization was the critical factor to getting virtualization into the datacenter, then virtualizing storage became the next logical step.
Not a Flash in the Pan
The article discussed the various features of current storage offerings, then defined storage virtualization, including popular options like switched-based, appliance-based and array-based. The article concluded with a look at some of the big players in the space, along with a number of up-and-comers.
The first thing that struck me about this article was what was missing: no flash storage!
Of course, it didn't really exist as a category then; it's been in just the last couple of years that flash has become entrenched in datacenters. There's basically no action in the HDD market anymore, which makes sense: flash is approaching (or has attained) mechanical storage costs, so why would a business even consider it?
I didn't realize how big flash storage had grown until I went to VMworld 2014, shortly after I'd been moved back to this magazine after more than five years away, covering software development. The number of storage vendors on the show floor dwarfed pretty much everything else, and they were all talking up flash.
Another major movement has affected storage, which the article couldn't have foreseen: the rise of software-defined storage, or SDS. SDS makes all storage resources appear as one pool of storage, available to applications. SDS is often combined with all-flash storage in hyperconverged appliances.
The ability of these two (relatively) new technologies -- flash storage and SDS -- to make storage faster and more efficient has had a huge impact on the industry. For one thing, it may help speed the spread of virtual desktop infrastructure (although not everyone agrees.) It will certainly be critical in the increasing adoption of cloud computing, both in the private and public realms.
It took many years after this article was written to get to that point, however, giving this article a lot of relevance in the intervening time.
Posted by Keith Ward on 07/14/2015 at 8:10 AM0 comments
More in This Series
Continuing our stroll down virtual memory lane by doing an analysis of Issue No. 1 of Virtualization Review, we come to Citrix. It was part of our cover story, an analysis of the leading virtualization players. It was titled, "Citrix Aims High," and it was written by Doug Barney.
One focus of the article was where Citrix goes with XenSource, which it had bought the previous year. XenSource, of course, became XenServer. Peter Levine, then with Citrix, was quoted as saying that the Xen approach was more efficient than VMware's: "The Xen hypervisor is a very lightweight infrastructure compared to what ESX has, just because of the design points and the assumption that there would be hardware assist under the covers."
XenServer Falls Behind
Whether he was correct or not, any advantages XenServer had didn't equate to success. Citrix released v6.5 of the product earlier this year, but it's safe to say XenServer has lost ground over the intervening time to both VMware ESXi and Microsoft Hyper-V. It hasn't been updated as regularly, or been enhanced the way the other two have.
In Gartner Inc.'s last "Magic Quadrant" for server virtualization companies, Citrix dropped from the Visionaries quadrant, where it was in 2013, to Niche Players. This followed a similar demotion the previous year, when it was dumped from the Leaders category to Visionaries. (Note that Gartner hasn't yet released its 2015 Magic Quadrant yet; if tradition holds, it will be out very soon.) The only vendors in the Leaders quadrant were VMware and Microsoft.
Here's how Gartner summarized XenServer: "For server virtualization alone, Citrix's reduction in XenServer functionality (at the same time that Microsoft and VMware are expanding) and clear strategic positioning will make XenServer an unlikely competitor."
VDI to the Rescue
A virtualization area that Citrix has done better in is virtual desktop infrastructure, or VDI. The story mentioned this in passing, because it wasn't a big focus at the time; it has, however, turned out to be much more important to the Citrix bottom line. "XenDesktop is a new product that sends Windows images out to a virtual desktop. It uses some Presentation Server concepts, but it's not the same product," Levine said.
Barney added this dash of analysis: "Competition for the desktop is heating up. VMware is also moving into the application virtualization space with its acquisition of Thinstall." Here we are, seven years later, and it's still the same two companies, duking it out for VDI supremacy. XenDesktop and XenApp are both solid, popular products that compete well with VMware Horizon.
One final note: As part of the article, we interviewed each company's executive responsible for virtualization: CEO Diane Greene of VMware; Bob Muglia, then head of the Microsoft Server and Tools Business; and Citrix CEO Mark Templeton. Templeton is the only one left; he announced his retirement in early 2014, only to change his mind the following June. He said in the interview that ran with the article that purchasing XenSource was a no-brainer for Citrix:
"It became obvious to us that virtual infrastructure was a core technology for us; we needed to be able to really bake it into our products for application delivery, and owning it was important. When we looked in the marketplace, though there were a few options, XenSource was by far the logical option for us to take."
Posted by Keith Ward on 07/10/2015 at 8:36 AM0 comments
There's a bit of a blog war going on between two prominent virtualization companies, which highlights some trends in the industry.
The combatants are VMware, the leader in the space, and Nutanix, a maker of hyperconverged appliances that combine all aspects of virtualization -- compute, storage, networking, management and so on -- in one box.
VMware, of course, helped create the industry and has long been the most dominant force. Nutanix, which got its start in 2009, is a much smaller company, but in its narrow niche of hyperconverged appliances, it's as important and influential as VMware. It's considered the leader in that area. (It's also important to note here that Nutanix sells a lot of vSphere in its products, making the companies partners, as well as competitors.)
There are two primary opponents here: Chuck Hollis, chief strategist, VMware Storage and Availability Business Unit (and former CTO for parent company EMC). He blogs regularly at Virtual Blocks, the VMware storage blog. In the other corner is Lukas Lundell, global director of Solutions and Performance Engineering at Nutanix. He blogs a lot for the company.
First Shots Fired
As far as I can tell, the spat started when Hollis began a series of blogs (now four in all) that compared performance testing between Nutanix and VMware, contending that VMware wins out in all situations. He's also stated that VMware is significantly cheaper to use than Nutanix. By my count, that's five blog posts that take direct aim at Nutanix.
Nutanix's Lundell has fired back, and his main response is on his personal blog. It has the provocative title "Nutanix vs. vSan Price and Performance — Part 4 (Or Why Chuck and EMC Are Deathly Afraid of Us)." So, nothing controversial there, huh?
The title is one of the most interesting parts of it, and it appears that Lundell has a point: Hollis has spent a lot of energy targeting a much smaller company that, on the surface, appears to be no real threat to VMware. There are numerous hyperconvergence vendors out there, including SimpliVity, Scale Computing, Maxta, Atlantis and others; included in that list, of course, is VMware's own EVO:RAIL. Because Nutanix sales can't compare in any way to VMware sales, why all the energy spent?
Lundell thinks he knows why: "I also didn't realize we'd attract so many mortal enemies bent on our destruction along the way. This is the cost of disrupting an industry filled with entrenched interests," he writes. While I'd chalk this up to unfortunate hyperbole -- trying to point out the differences between your products and the competition, as Hollis is doing, is hardly the same as being "bent on destruction" -- it is true that VMware is entrenched. It's also struggling in some areas, including (possibly) EVO:RAIL sales, and public cloud.
Of course, one person's "entrenched" is another person's "very successful," and VMware fits both definitions. It continues to make more money every quarter, and last year it surpassed $6 billion in revenue for the first time. VMware isn't going anywhere for the foreseeable future, and appears healthy by most metrics. Given that, why all the focus on Nutanix?
Hollis spelled it out in his response to Lundell's post. He says, "… nobody here is afraid [of Nutanix] -- especially me. I know that makes for a colorful storyline, but it's just not true. What we do see is a small startup making ridiculous claims … which can confuse some customers. That's not good."
Testing ... 1, 2, 3 ... Testing
Much of the imbroglio has to do with testing. Hollins says that VMware's objective testing conclusively demonstrates its claims of better performance at a cheaper price. But, he points out several times, the Nutanix End User License Agreement (EULA) forbids publishing any test results of its products. Lundell says, basically, hogwash. "In the real world, the testing of these systems is not so simple," he says, and suggests a "real-world test methodology" that would include various types of tests.
And on it goes. Here are a couple of my takeaways from this back-and-forth:
1. VMware may indeed be afraid of Nutanix. Although Hollis rightly says that competitors often do make "ridiculous claims" about rivals (note that I'm not saying Nutanix is doing this; just that this sort of thing does happen, and frequently), he doesn't single out any other rivals -- just Nutanix. It may also be an indication of how important VMware feels the hyperconverged appliance market is about to become.
2. Lundell should seriously consider toning down his rhetoric. It doesn't help when one guy says a competitor "… will stop at nothing to steal, stifle and stall innovation to protect their monopoly. Lawyer fracases, intentional misdirection, blatant misinformation, and "brain-rapes". It's all on the table." I've never seen a discussion, ever, anywhere, on any subject, improved by taking these types of cheap shots; it only serves to lower the bar. Treating others with respect is much more likely to lead to a positive, helpful discussion for all.
3. Beware testing claims made by anyone. I've always felt that tests like this are of relatively little value to an admin. That's because the environment used for the test is 100 percent likely to not be your environment. Only you know what your needs are, what kind of infrastructure you have, what kind of workloads you're running, what your network is like, and so on.
If at all possible, perform the testing yourself, in your environment, with your workloads. Pretty much every vendor will offer a trial period, and you should take advantage of it. What works best for your datacenter may work worst for the company in the next building.
(Note that I did not interview either Hollins or Lundell for this article. My purpose was to highlight the discussion and offer a few thoughts on it. Both men state their cases eloquently.)
Posted by Keith Ward on 07/09/2015 at 4:14 AM0 comments
More in This Series
Continuing our stroll down virtual memory lane, we're going through some articles from the very first issue of Virtualization Review, from April 2008, getting a view of where the industry was then, and where it is now.
This time, we'll look at the article "Microsoft Reaches for the Virtual Sky," which was part of our cover package on the virtualization leaders of that day. In the story, written by me, the theme is "integration." Microsoft wasn't interested in being a virtualization vendor: It wanted to be an enterprise OS vendor that used virtualization as one more tool to help sell Windows Server.
A Consistent Strategy
Sounds familiar, doesn't it? Microsoft still has pretty much the same game plan. Certainly, it's moved forward in a number of areas, but its virtualization offerings are meant to bolster Windows Server. It's never wanted to be VMware, and rarely competes directly (although that's happening more and more as the company goes after areas like cloud computing). Virtualization has remained an ingredient in the pie, but not the pie itself.
As an example of how embryonic Microsoft's virtualization strategy was, Hyper-V wasn't even out yet -- it would be released a few months later, with Windows Server 2008 (how many remember that Hyper-V was code-named "Viridian"?). It was Microsoft's first bare-metal hypervisor, and was a bit late to the party, with VMware ESX and Xen having already been out for some time.
A Bad Omen
And even with its tardy arrival, it didn't arrive full-featured: "Some technologies that were scheduled to debut in Hyper-V were pulled last year so that Hyper-V could be released with Windows Server 2008 (WS2008). Functionality like live migration, which allows a VM to be moved from one physical machine to another with no downtime, and several other key features were dropped." I remember that being a big deal at the time; for such an important product to be stripped down just to meet a shipping date? There were more than a few who thought this might portend bad things for Hyper-V.
Those fears proved to be overblown. In the intervening years, Hyper-V has been substantially beefed up, including the addition of its enterprise management tool, System Center Virtual Machine Manager, and has overtaken Xen to be the No. 2 hypervisor in enterprises, after ESXi. But it was definitely a rocky start, and more than a few folks wondered if Hyper-V would gain any traction, given Microsoft's handling of its release.
A Gartner Inc. analyst had a more optimistic take, and has proven to be right. From the article:
"I think [Hyper-V is] going to be utilized at a fairly high rate. Clients are waiting and planning on testing it, or have made the assumption they'll be implementing it," says Jeff Hewitt, research vice president at Gartner Inc., who specializes in server virtualization. "Quite a few people are going to be evaluating it, and implementing it once it's available," he predicts.
Enterprises have grown comfortable with Hyper-V. The majority of shops these days are multi-hypervisor, as admins use what works best for a given workload. Although Microsoft is still not thought of in the way VMware is when it comes to virtualization, the company has staked out a comfortable spot, and its modest goals for virtualization have worked out -- in large measure because they were modest from the beginning.
Posted by Keith Ward on 07/06/2015 at 8:44 AM0 comments
More In This Series
I recently started a series of posts on the first-ever issue of Virtualization Review magazine, looking back at the state of the industry then, vs. what it is now. This time, I'll examine our cover story, which was an overview of the three dominant players in virtualization in 2008: VMware, Microsoft and Citrix. And wow, are there some interesting nuggets to pull out of that feature.
The first article, written by me, was called "View from Above." It was about how VMware became the biggest company in this fledgling industry, and where it might go from there.
Even way back in early 2008, VMware was seen in some ways as a threat. Here's what I wrote then: "Some of the chief complaints about the company are the proprietary nature of its software, the high cost of its products and a fear from some competitors that the company is becoming too powerful -- all of which comes with being 'King of the Hill.'"
These complaints are still being heard today. By comparison, for example, Microsoft has become much more open source-friendly than VMware. And many still lament that VMware is the most expensive company out there for strictly virtualization-related products like vCenter. The counter-argument is that despite the expense of its products, VMware still has a commanding market share in those areas, so those offerings seem to be the ones that companies want most.
Another tidbit from the article is that VMware sales in Q4 2007 were $412 million; the similar quarter last year saw revenues of $1.7 billion. That comparison also shows the growing saturation of server virtualization, which continues to be VMware's cash cow; the 2007 number reflected 80 percent year-over-year growth, while the Q4 2014 income was an increase of 16 percent year-over-year.
Related to that, I wrote: "But VMware, perhaps sensing that hypervisors may be moving to the realm of commoditization, is quickly adding to its portfolio of non-hypervisor-based offerings." The company has continued to do this at a very steady pace, not settling for just being a server virtualization business. Its offerings have gone beyond just new virtualization-related wares, too: it does a brisk business in mobile device management with AirWatch, for example, and continues to try and find its way in cloud computing. It's also worth noting that in 2014, VMware surpassed $6 billion in sales for the first time.
Keeping Ahead of the Competition
I also discussed VMware's competitors, and whether it could keep ahead of them to remain at the top. Dan Kusnetzky of the Kusnetzky Group (and now a columnist for Virtualization Review), said it faced real challenges: "As far as their share of the industry's mindshare," comments Kusnetzky, VMware is the "dominant supplier and able to charge premium prices; but that's under attack from Microsoft on one side and the open source players on the other."
Indeed, there are still plenty of competitors out there, and they still include Microsoft, Citrix and various open source companies like Red Hat. But thus far, VMware has managed to hold them off, mostly by continuing to innovate in its core areas and careful expansion into others.
Not Easy Being Greene
This article was accompanied by an interview with co-founder and then-CEO Diane Greene. Ironically, it would be one of the last interviews she did as CEO: she was ousted just three months after this issue hit, and replaced by ex-Microsoft exec Paul Maritz. She was a true visionary, and did a lot to make VMware a dominant company. She was also one of the few female CEOs in the industry, a situation that sadly continues to this day.
Greene hasn't been much in the public eye since then, although she appears to still be active in technology startups while keeping an extremely low profile.
Posted by Keith Ward on 06/30/2015 at 8:54 AM0 comments
More in This Series
Earlier this week I stumbled across the very first issue of Virtualization Review magazine. It was the March/April 2008 issue, and it was a new thing in the industry: the first print publication to cover the emerging field of virtualization.
What's a Hypervisor?
When I was first asked about being editor of a magazine covering virtualization (in 2007, many months before we produced the first issue), I remember thinking "OK, but first let me find out what the heck that is." I knew the word "hypervisor," and knew that it sort of had something to do with abstraction. I was also familiar with Parallels software, which I had on a MacBook Pro. I figured it was similar technology.
If it sounds like I was pretty clueless about virtualization, you're right. I still thought the "one server, one app" paradigm would dominate for years going forward. I knew that such an arrangement was an inefficient way to operate, but hey, it had worked for a long time; why mess with a good thing?
Yep, things have changed in the seven years since that first issue went out the door. Looking through the magazine brought back a lot of memories, and has also given me a strong sense of how far the industry's come since then. So I thought it would be fun to go through the issue with you, making some observations and comments about the past, present and future of virtualization, using that first issue as a guide.
Taking Aim at VMware
Let's start with the lead story in the UpFront section, which was dedicated to news and analysis. Here's the headline: Analysis: Taking Aim at VMware. The deck reads: With the release of Hyper-V, is VMware in danger of becoming the new Netscape? With the benefit of hindsight, that seems like a pretty silly question to ask.
At the time, though, it wasn't silly at all. Microsoft was just a few months away from including the first version of Hyper-V in Windows Server 2008. Hyper-V would be a free offering, and more than a few people were wondering if we were about to see Netscape 2.0 play out all over again: Would Microsoft's bundling of a new technology for free in Windows push aside the current market leader in that space?
As Analyst Al Gillen of IDC said in the article, "Microsoft has to be careful about how it bundles products together to not run afoul of antitrust regulations."
Still Leader of the Pack
Of course, that didn't happen, and over the years, VMware has remained comfortably the leader in virtualization, even though Microsoft has made some inroads. In the article, Gillen made a statement that's proven to be prophetic: "Eventually, the hypervisor will be kind of generic, so that's not where the battles are going to be fought," he said, and he was right. VMware has maintained its leadership, and keeps making vSphere better and better -- witness upgrades in vSphere 6.0 like Virtual Volumes, long-distance vMotion and improved Fault Tolerance. It's staying ahead of Hyper-V, although Hyper-V has also been substantially upgraded over the years.
However, while VMware remains in the lead when it comes to server virtualization, it hasn't fared as well in the current big-growth category, that of cloud computing. Although its vCloud Air product is well regarded, it hasn't moved much beyond the private cloud realm, while Microsoft Azure is firmly ensconced in the No. 2 spot behind Amazon Web Services (AWS), and is even challenging AWS in some circles.
More to come in future installments, as we continue our stroll down virtual memory lane.
Posted by Keith Ward on 06/26/2015 at 9:17 AM0 comments