For those with a handy QR Code Reader, read no further! Here’s a beautiful piece of motel art for Brian that says it all:
It must be that time of the month over at BrianMadden.com. That is, the time of the month when click-through rates for advertisers are pretty low because – let’s face it – it’s difficult for Brian come up with new stuff every day. So he’s figured out a new strategy. First, rely on being sensational, and let the readers contribute the real value. Second, every once in a while pour withering scorn on some vendor (yippee, today it’s Citrix) from his tech pulpit in the hope that we will respond and drive more clicks to his site.
Brian compares Citrix products to Motel Art – generic, undifferentiated and designed not to irk Microsoft. I must say I love the analogy, and Brian makes a couple of valuable observations, but I think it’s fair to say that his comments indicate that he has no clue whatsoever about building products that customers value. Below are a few of his more ridiculous pronouncements, with some questions for him in return:
“Citrix makes a client hypervisor. It’s good, but not so good that too many people actually use it which would piss off Microsoft. (Appearance of innovation? Check! Not pissing off Microsoft? Check!)”
XenClient is the industry’s only enterprise-class type 1 client hypervisor, period. It has succeeded in delivering on key requirements for highly secure deployments of off-line desktop virtualization with an excellent HCL for a first release. It is also the industry’s first ever open-source based bare-metal client platform to have successfully tackled the daunting challenges of virtualizing devices such as GPUs, that were never intended to be virtualized, to deliver incredibly good performance given the constraints of modern laptops (battery life, memory, networking and other issues). XenClient was put together by an incredibly small team of black belt engineers, very fast, and with very high quality. Meanwhile, VMWare’s project, CVP, utterly and miserably failed, and there has been zero proof that any other vendor has made credible progress in this challenging area. It just turns out that the use case that we targeted for XenClient (concurrent consumer and enterprise environments, with absolute isolation and high security) is one in great demand and one that cannot be served by any native Windows environment, and so Microsoft is positively disposed towards XenClient. Microsoft is a great partner of Citrix, and we respect their business, and they ours. What’s the problem with that? And did you see anyone at Citrix say “XenClient 1.0 is the first and last you’ll see of this product.”?
Citrix claims to “embrace and extend” RemoteFX, but they only do so in the exact ways that make Microsoft happy. So Citrix only supports RemoteFX on XenDesktop (no XenApp), only to Windows 7 clients (no iPad/WinXP/Android/etc.), only connecting to Windows 7 SP1 hosts, and only on Hyper-V. (So check, check, check, and check for not rocking that Microsoft boat.)
Since you are clearly no longer an engineer, you probably don’t know how long it took for those of us interested in using and extending RFx to actually get working code from Microsoft. As it is, we are lucky to have made the considerable progress we’ve managed thus far, in delivering RFx from XenDesktop. To date we have seen no API support for RFx from Microsoft, for Windows TS. As soon as they provide it, it is our full intention to support RFx across our entire product set, but engineering bandwidth is finite, and moreover Windows 2K8R2SP1 footprint at customers thus far is tiny, and its install base on servers with GPUs is miniscule. I will look forward to your glowing praise as we deliver RFx more broadly.
Citrix has XenServer, but for both Geek Week and my VDI experiment, they wanted me to use Hyper-V. Also they share all their best features with Microsoft via Citrix Essentials for Hyper-V. I love how they don’t even include the word “XenServer” in the “Essentials for Hyper-V” product.
Where to start?
- XenServer is awesome, and the majority of new sales of XenDesktop are now on XenServer, because we dramatically reduce the TCO and enable key requirements for security of virtual desktops to be met. This is not possible on any other virtualization platform, because we have focused heavily on optimizing image management, security and storage for desktop virtualization.
- Hyper-V is a great hypervisor – fast, great density, bulletproof. Microsoft doesn’t offer storage (or network) optimization of any form, least of all for VDI. Where the customer chooses XenServer tends to be because of our integrated virtualization of server, networking and storage in a powerful way that dramatically cuts TCO for VDI. Neither Citrix nor Microsoft charges for our hypervisor. Our shared competitor charges a phenomenal amount for a product that doesn’t deliver any benefits at all versus either Hyper-V or XenServer.
- Again: we don’t monetize XS, and Microsoft doesn’t monetize Hyper-V. So we care simply that the customer has a choice, and that the choice between one or the other is fully vetted by the customer. Recently we have been driving to achieve density and scale proof points for XenDesktop on Hyper-V given the release of W2K8R2SP1, and it was presumably in that spirit that Citrix engaged with you to try XenDesktop on Hyper-V. I will ask our field teams to remember in future that any mention to you of “multi-vendor support” will be transliterated to mean “XenServer is lousy and you shouldn’t use it” in your scrambled mind.
Citrix maintains a XenApp streaming capability, but you have to dig to find it. And even on that page, they finish with, “Hey, in XenApp 6 you can publish App-V packages directly from XenApp.”
The primary use of XenApp streaming has been to deliver client apps onto XenApp servers. Our feature set supported 64b apps well ahead of Microsoft App-V. It is widely used by our customers. It’s always been “just another feature” and not something we specifically set out to monetize. Customers don’t often use it to stream apps to PCs – for that they purchase App-V. Now, Microsoft monetizes App-V, and moreover it supports 64b applications. It’s pretty obvious that when a product is monetized as a part of the platform by Microsoft, it will likely become the dominant technology in use by customers. For Citrix it is an absolute don’t care. Our streaming works fantastically. Use it if you will. If you don’t, then use the product that Microsoft sold you. Does that mean it’s Motel Art? No. The technology is mature, and works very well. It’s just a feature.
While VMware is understands that the world is moving away from Windows by buying SpringSource, establishing Horizon at easy-to-reach price points, etc., Citrix is still firmly focused on a Windows-only (i.e. “Microsoft-only”) world. Sure, Citrix has Receiver clients for Android, iOS, Mac, Chrome, etc., but those only exist to connect back to your Windows apps.
Spring is a great (OSS) dev environment for enterprise Java apps. It is not a PaaS platform as VMware seems to think. Spring based apps can run in many frameworks, such as Makara, GigaSpaces, and many others. Spring does not address any deployment challenges such as security or elastic scaling, and dealing with the networking issues required to “enterprise-up” a web-based application, whether running at home or in the cloud. We have done all of that work, for Spring (eg: see here), or .NET, or PHP or whatever else you happen to want to use to write your app. Citrix has the world’s leading portfolio of networking products for any app run in the cloud, and yes, we monetize it. The market sets the price, and we win more than our fair share of bids. Does VMware have any way of delivering an app from the cloud with an Enterprise-class SLA, meeting compliance requirements, or triggering elastic auto-scaling? No. All it has is a staggeringly expensive compute virtualization platform that has an over-reliance on expensive, outmoded infrastructure technologies, and a pathetic notion of VM compatibility as a defensible play. What did they give you to drink with breakfast?
Spring Mobile is a hail mary PR spin on a poorly judged acquisition. It remains to be seen whether it will have any relevance in next gen mobile app development. While I might one day write an Android mobile app in Spring, I could also write it in Eclipse or any other of the hundreds of app development environments that have sprung up for mobile devices. Did you happen to notice that Spring can’t push out apps for Apple devices, or that Java != Cocoa? Ever think about the fact that there are extremely important infrastructure platform components, including caching, crypto, acceleration, app firewalling and end-to-end data security that are required below the app dev environment? Do you realize what vFabric really is?
Project Horizon should be Project Mirage. As soon as you get close to it, it retreats off into the distance with vague promises of relevance in the future. Thus far I’ve seen nothing remarkable in the offering. A few simple API tricks and partnerships with app vendors, and a rather ghastly looking fake desktop on my iPad that certainly does not inspire allegiance and is likely to antagonize Apple. Finally (and quite remarkably, given your thoroughness) you fail to remark on the fact that Citrix Receiver is entirely free and supported on just about every client, when VMware has only just managed to get PCoIP support on an iPad, assuming you could get it over the WAN, through a firewall and to the device.
I could go on, but I won’t. This is the first time I’d say Brian has truly demonstrated cluelessness. Or perhaps it’s his usual genius, but in disguise – possibly millions of new clicks are winging their way to his site as I type.
Hey, I have an idea: Let’s debate this face to face on stage at Citrix Synergy. Feel free to bring along @Herrod or any other apologist for VMware – I personally commit to paying your and their entrance fees from my own pocket.