Is UPD now FSLogix? Taking a look at the FSLogix acquisition by Microsoft.

Perfect timing I guess. A couple weeks after I released the whitepaper I wrote showing how UPD compared to FSLogix, Microsoft decides to open its wallet and acquires FSLogix. I am sure someone at Microsoft did read the whitepaper and understood that UPD needed a revamp and that it would probably take them a long time to fix it than opening their wallet. Very happy to see this happening. With that in mind, let’s take a look at what this potentially means to everyone in the industry.

I am not going to discuss the existing UPD limitations and how FSLogix can be used to complement it or to fully replace it. You can read all about that on this blog post.

The question now that many are asking themselves is simple: will this be part of a cloud-only offering, like Windows Virtual Desktops is as of today? The answer, no one really knows for sure. Probably, not even Microsoft.

The main thing is, UPD, even though it is a much better solution than traditional roaming profiles, still suffers from many issues, no matter if you are hosting your solution in the cloud or not. At the end of the day, you are still accessing a Windows OS and given how it works, a profile is always required (even if it is a local one).

If UPD 2.0 (that is how I will call the FSLogix offering, now under the Microsoft umbrella) does become what FSLogix is and more, it makes no sense to tie it to a cloud-only offering. The reason for that is simple. First of all, having to deal with two different solutions for on-premises and cloud based deployments. Considering many are still fully on-premises and some are in a transition mode (one that may take years), forcing customers to have to deal with two completely different solutions, especially when in a hybrid deployment, could lead to a terrible end-user experience, where things do not work smoothly regardless of its location.

And that is something that Microsoft is trying to avoid at all costs at this stage. If the plan is to turn Microsoft into an utility company, where you pay your monthly bill exactly the same way you do with your cable, natural gas and hydro, it has to behave exactly the same way as everyone is used today. To the point that no one can tell the difference where it runs or how it behaves. Once that is the case, almost certainly a transition to the cloud will be just a natural evolution of an on-premises environment. Simple, effortless and more than that, painless.

Making such solution a cloud-0nly offering creates this gap between what is there right now and what will be there in the future, simply creating push back from customers, instead of promoting adoption. Reason why I do believe that Microsoft will do something regarding WVD, making it available on-premises. To simply make the transition to a cloud-hosted WVD simple for anyone using WVD on-premises.

Yes, luring customers, instead of forcing them.

Now it is worth mentioning that Office 365 is far from being a Windows only offering. Many use it from mobile devices running iOS and Android and of course from non-Windows desktops. With that in mind, to make the Office 365 experience the same no matter where/how you use it, Microsoft has to fix more than profiles. As per my twitter, the main one that comes to mind is printing. I do remember a session I attended during the first ever BriForum, in 2005. Yes, thirteen years ago. And guess what? The printing landscape is as bad as it was back in that day.

So Microsoft, please keep your wallet open and get Tricerat as well. That will give us, Office 365 users, a true world class experience, no matter where we are and what we use.

You know, like a true utility company, that does not care if I have a Vizio or a Samsung TV.

It just works.

CR

540 total views, 1 views today

The rise of VHD-Based Profiles. And the marketeers.

I decided to write this post due to the fact I am getting tired of marketing people in general, always attempting to sell you something you may not need and worse than that, trying to spread fear all over the industry about other solutions. Before going ahead, let me make one thing clear: I truly believe every product out there in the EUC does have a reason to exist, beyond making money. They do address a particular need and certainly have their value and merit.

Now, leaving the marketing bullshit behind, that does not mean any of these products are the silver bullet, the one solution that will solve all your problems, with zero side effects. If you ask any vendor what the drawbacks are with their product and they have no answer to that, please, do yourself a favor and run away. Every single product has drawbacks and issues. Period. The key thing is understand these and how you can minimize or eliminate them (with potentially another product to complement the first one).

With that in mind, let’s have a quick chat about VHD-Based profiles, what seems to be the hotcake these days. If you are not aware of, Microsoft introduced User Profile Disks (UPDs) back in 2012 with Windows Server 2012. Yes, not even R2. That means whatever this is, it is SIX DAMN YEARS OLD. Got that? Six years in computer years is like 120 human years. Just to put in perspective (I do know you talk about dog years at home, so let me help you making things simpler) how damn old this is.

The idea behind UPD is very simple. The C:\Users\%USERNAME% folder gets pointed to a Virtual Hard Disk (VHD), a single file, sitting somewhere. How big can it be? No idea on the limits but I have used them set at 20, 40 GB without issues. That means every user will get a file that can grow up to whatever it was set to (i.e. 20GB) and that file will get mounted and linked to the user’s own C:\Users\%USERNAME% folder.

Right off the bat you can see that if you have let’s say 40 users connected to your RDS Session Host (XenApp for Citrix people), each user will have a profile folder with 20GB. That means 800GB for user data. Note the C: drive on the server is usually 60-100GB in size. This is possible as it is just a mount point. You are not using disk space off the C: drive but you are still able to have users with profiles that could be potentially bigger than the server drive itself. Nothing magical here and more than that and one more time, SIX years old. But marketing people want to make you believe they are now selling magical software that can magically make your local drive grow like Godzilla. Nope.

As it is a single file, when the user logs in, there is no need to download anything to the server drive. The mount point is established and you are done. Does not matter if the UPD has 2GB or 200GB. Logon time will be the same and as it is just a mount, it will be much quicker than using traditional solutions (i.e. roaming profiles). Here we have the marketing geniuses again, trying to make you believe you are buying an amazing technology that makes your logons much faster now that you are riding on Unicorns. I can make logons faster too and I do not even work at marketing, or have unicorns, just for the record.

Back in December I presented at the Citrix User Group in Israel, exactly about this topic. I showed it live on stage, two completely different solutions (Citrix XenApp and Parallels RAS) up and running, where the same user had UPD enabled. When he logged in to Citrix and did whatever he wanted and logged off, once he logged back in but now through Parallels, all the stuff he had done on XenApp was there on RAS. To add a nice twist to the whole thing, I had the Parallels environment on Azure. That means I was replicating UPDs ON THE FLY, LIVE, between an on-premises solution running Citrix and a cloud-based one running Parallels, for all my users. As you guessed, yes, a completely agnostic solution that does NOT care which product you have and where it is running. And the best part of all this, FREE. Yes, this is part of the Windows Server feature set. No matter if using VMware Horizon, Citrix XenApp or XenDesktop, Microsoft RDS or Parallels RAS, this works out-of-the-box and with all of them.

That said, is UPD perfect? Not at all. It has its limitations (i.e. cannot be mounted twice) like anything else. But it is certainly a powerful solution that is worth investigating and testing. Thing is, many companies realized that a long time ago and now sell their own solutions that in a nutshell use the EXACT same principle. Mount the user profile to a VHD and name it profile container, profile disk or whatever they want to call it. Are they better than UPD? For certain use cases, of course they are! FSLogix for example allows you to mount the VHD multiple times and does use its own filter drive that allows apps like OneDrive for Business to work under RDS. If you do need something like that, sure, take a look at FSLogix (as far as I know, Liquidware Labs does have a similar product, that addresses similar issues – may not address the SAME issues).

The lesson here is simple. UPD, profile containers, VHD-based profiles or whatever you want to call this, is not a new thing. It has been around for a long time. It is not something new or magical as many of these vendors try to make you believe. And what pisses me off the most is the simple fact they try to make you and the industry believe that UPD should never be used, that it sucks and so on, what goes completely against what I think that is always to use the RIGHT TOOL for the RIGHT JOB. Some vendors like FSLogix even got pissed at me with the whole UPD story. Seriously.

For the companies out there, stick to honest marketing and sales and educate your customers and the industry properly, clearly showing what can be achieved with the out-of-the-box solutions and what you bring on top of that.

For you, readers, at the end of the day, it is up to you to decide which tool you need and if you feel like using a screwdriver to put down some nails, go for it. After all, as my wife says, “Why do you have a Lamborghini to do your groceries?” and to that, I have no answer. But do not make the same mistake as I made and make sure you get a hammer to handle some nails.

Cheers.

CR

692 total views, no views today

Nirvana Phone. Is the Motorola Atrix the answer?

Today Motorola announced their new flagship phone, the Motorola Atrix. Citrix right after announced they are shipping the Citrix Receiver on this device and Chris Fleck also blogged about this phone as being the first true ‘Nirvana Phone’, a term they coined years ago for a phone that one day would be the perfect replacement for the road warriors out there.

As this info was under NDA, before the announcement I was discussing which phone this would be (as none of the CTPs had that information) and I mentioned before the announcement that I was certain it was the Atrix, what was later confirmed. 🙂

So here is my take on the Atrix and on the whole Nirvana phone concept and how I see this working one day.

First it will take more than manufacturers to make this idea work. The goal is to have a phone and from that be able to do almost everything you need during the day. Sure people have different needs and for that reason no phone will ever be the silver bullet that will work for every single person out there.

Secondly, if I need to carry ANYTHING, it defeats the purpose IMHO. For example, as of today even with the Atrix, you will need to travel with a bluetooth keyboard (and potentially a mouse), a dock, cables to hook this up to a TV/Monitor and so on. Well if that is the case I will carry my DAMN LAPTOP!

Of course if hotels for example would start offering bluetooth keyboards/mice combos even for a small rental fee per day, that would help a LOT this idea. I personally would pay US$ 4 a day for this. After all, the target market for this is probably someone that is always on the road on business so US$ 4 a day extra on expenses will not create another breakdown of the US economy. Even with the hotels jumping into this idea, unless they also provide docks and cables you would still need to carry something.

That brings us to Apple and their iOS. Not many people dug into AirPlay and what this could mean for this concept. If Apple opens up AirPlay to third parties AND work with other manufacturers to embed this into TVs, Monitors, Receivers, etc this would make an iOS device the true Nirvana thing. Just being able to walk into a hotel room and from your pocket send the video-out of the Citrix Receiver to the TV and wirelessly use a keyboard/mouse would be much better than having a dock and cables, even hotel provided ones (what I do think would help a lot this ‘Nirvana Phone’  concept). An iPad in this case would be the perfect thing to carry as it has a large enough screen for the local apps and processing power to run some very decent ones (what would mean companies writing iOS apps that tie into their backend systems/databases) when offline and when in a hotel room, use the Citrix Receiver/AirPlay to really shine as a Thin Client.

Regarding the Atrix, I still think it suffers from the same issues as any other Android device: giving the power back to the carriers. As an example, the Android device I have, bought 18 months ago from our loved carrier in Canada, Robbers (ok, it is Rogers), came with Android 1.6 and the freaking bastards over there refuse to release ANY updates for this particular phone. So once I rooted the damn thing and loaded Android 2.2, things got better. But this is NOT supported by the carrier and more than that, it is not really something you would give to most users to do. That is the main reason why I think Android sucks.

Add to that the fragmentation now created by Motorola, LG, Sony, Samsung and so on where an Android application that runs on a device like the Atrix will not work the same way on an HTC phone with their ‘Sense’ interface. In that respect, controlling the OS proved to be the right thing to do, at least from taking the power off the carriers and giving it back to the users. At the end it is up to me to decide if I do want to run the latest and greatest OS on the device I BOUGHT, not to the carrier. Their answer to that is ‘just buy the new model and get into another three year contract with us’. Sure, they want to milk the cow. Yes, you are the cow.

For the Apple bashers/haters out there, I do accept the fact my old iPhone 2G cannot run iOS 4. It is the same reason why your Pentium MMX laptop cannot run Windows 7. This market is used to that. But not intentionally preventing your Intel i5 machine from running Windows 7 just because it shipped with Windows XP.

So back to the whole Nirvana phone topic and resuming: I like the idea but I do think the Atrix is far from being the ‘One’ (hint to Motorola – name the next one Motorola ‘Neo’) and without help from other businesses like hotels, there will always be a drawback for this concept to really take off. And finally in my mind Apple is the one that can pull this off, given the traction iTunes will bring to AirPlay enabled devices like TVs and Monitors, really making things easier and better for us consumers.

I guess time to start a keyboard/mouse rental business that will partner with someone like Hilton…

CR

1,589 total views, no views today

Intel buys McAfee. Does it matter?

As you all have heard today, Intel bought McAfee. And will pay well for it.
So the question now, at least from my end, is how this can possibly affect or help the virtualization market, the one you and I live and breath on.
First of all, we all know for VDI to take off as a mainstream solution (what it is NOT at this day in 2010 – and will not be for a long time) it must get cheaper. By cheaper it means being able to cram more instances per server. This can be achieved in several ways like using the latest, greatest and fastest CPUs you can get and with as many cores as possible (and of course using quick ass disk subsystems like Fusion-io, caching/dedup like iLIO, using tons of RAM, etc).

The point is on the CPU space there is nothing that prevents other vendors like AMD to get to the level Intel is at. In many cases in the past AMD actually delivered better silicon than Intel. So Intel needs a way to differentiate itself from their competitors. Bringing stuff like AV closer to the HW is one way of doing this. Good for Intel.

And of course getting this OUT of the VMs will for sure increase scalability. That was the reason why McAfee and others were coming up with appliances and lightweight agents (to run on the VM) to offload all that work outside the virtual environment.

The main question now is really how Intel will pull this off by not being a software company really. How they will get McAfee going. Of course I think it is just way too soon for any analyst to say anything about this. Historically Intel has not managed acquisitions like this well but they were never on such scale and with such reach like McAfee has (good or bad, they do have customers and a name in the industry, especially since that .DAT file fiasco that screwed up more computers in a day than any virus they were trying to protect).

In the near future I do not expect to see anything embedded at the HW level. This is for sure something that will come way down the road as you need to come up with something that can be leveraged to anything running on top of that HW. This means you either change the OS that will be running to benefit from these new HW extensions (like vendors did when using the virtualization components exposed by Intel and AMD on their CPUs) or, in this particular AV case, you get an agent running on the OS/VM. Not an easy task to do considering the amount of hypervisors and OSs now available.

That leads us to a very important thing. To minimize this and make things much easier, would not make sense for Intel to grab a Hypervisor vendor now? Given the three main players on this space now, VMWare, Microsoft and Citrix, I am sure the low hanging fruit here is Citrix and I even wrote about this ages ago on the post ‘Intel buys Citrix’. 

This would give Intel a huge advantage over any other company in the Virtual Wintel echosystem. Controlling the CPU, the Hypervisor that runs on it and extra features like AV, would give you the ultimate virtualization platform, where your solution runs better or has more features than anyone else. Example? All the HW fancy features are only exposed to your own hypervisor (like Microsoft is doing with RemoteFX, only available to Hyper-V hosts) and of course your hypervisor will scale much better than the competitors as you own and know it all about the underlying HW platform. Then the next logical step would be to acquire a graphics company like NVidia (as AMD owns ATI) and leverage all that into the platform, exposing it to the virtualization layer. Then, buy a good storage vendor and a management/layering one and they are all set.

Sure such scenario could potentially bring a lot of issues to Intel from a legal perspective, as it did to Microsoft when it became what it is today. But certainly it would simplify the virtualization market a lot (and yes, I know, locking everyone into their platform – what may not be a terrible thing as Apple has shown the sceptics with their iOS echosystem).

The bottom line is this acquisition for sure will help the virtualization space in the long run (do not expect mystical benefits happening overnight with this acquisition) but I see it as just the tip of the iceberg of what is potentially coming down the road from them.

Feds, you better keep an eye on Intel.

CR

935 total views, 2 views today

RemoteFX First Impressions

As I did not have much time to test RemoteFX extensively, here are the first impressions of it and how we got it to work.

First of all, you MUST get a compatible video card. Not everything will work with Windows Server 2008 R2 with SP1 with Hyper-V, so you can get your Windows 7 VMs (with SP1 of course) working with RemoteFX.

I posted about it before. You can read the list of supported video cards here.

What did we get?

– HP desktop with a six-core AMD CPU and 8GB RAM.
– FirePro 5800 Video Card (also tried the unsupported Quadro FX 580 that by the way, does work too).

Initially I simply tested the Windows 7 VM connecting from the Hyper-V host itself but later got another Windows 7 SP1 box and used that one to connect to the VM.

Performance is decent I must say. I tried playing some Windows Media HD videos (make sure you disable multimedia redirection by using videoplaybackmode:i:0 in the .RDP file (save the RDP connection to the desktop and open it using Notepad). Also very important that you set the policy for RemoteFX (as I was not sure where to set it, I set it both on the client and on the VM itself). It is described here:

To set the experience index for connections using RemoteFX

  1. Log on to the client computer as a member of the local Administrators group.
    Click Start, and in the Search programs and files box, type gpedit.msc and then press ENTER.
  2. Navigate to Computer ConfigurationPoliciesAdministrative TemplatesWindows ComponentsRemote Desktop ServicesRemote Desktop Session HostRemote Session Environment.
  3. Double-click Set experience index for connections when using RemoteFX.
  4. Select the Enabled option.
  5. In the Screen capture rate (frames per second) box, click Highest (best quality), and then click OK.
  6. Restart the client computer.

The key thing to understand here is, why you may need RemoteFX. For example, during our tests, playing the WMV-HD tests, it used up to 30MBits so as you can see it is VERY bandwidth intensive. For comparison, running Google Earth in DirectX mode used around 9MBits. So basically the bandwidth will of course depend on the application being used. The same for how intensive CPU/GPU utilization will be.

I would expect applications like AutoCad to use way less bandwidth than something like WMV-HD and what we will be testing next is actually using RemoteFX over a typical home (cable/DSL) connection, simulated in our lab. By typical I mean a 10MBits down/1MBits up with 40-50ms latency and some packet loss probably in the 1% range (or a little more due to bursty loss). Given the first results we have seen, I am confident RemoteFX can indeed work over the WAN (at least bandwidth wise) depending on the applications.

Yes, before Brian Madden sends me a tweet or leave a comment here saying ‘MS says RemoteFX is LAN only’, I still want to make the point that IMHO, anything that is LAN only has its fate determined already. DOA. See my post about this here.

And still on the performance side, what we have seen in a nutshell is this: RemoteFX does work great BUT it is NOT the same as local. Simple things like Flip3D (using Windows key + Tab) are NOT as smooth as running them locally. Even Google Earth (that works just fine by the way) is NOT as smooth. But they both work and work fine, considering you are over RDP. For a BETA release we can expect it will be tweaked and improved even more before it hits the market.

As a sidenote, keep in mind there IS a bug on SP1 that throws a message on the RemoteFX event log about CPU encoding being used for ATI cards. It is a known issue and has been fixed apparently on later builds, what of course I have no access.  But for 1 VM testing like we did (I am after experience testing and not scalability – I will leave that to people with more time and resources on their hands like Ruben and Benny 🙂 ).

As soon as I have more results and some nice videos to show RemoteFX, I will post these here.

CR

4,472 total views, 1 views today

LAN only protocols for VDI. DOA?

As promised (I know, late) here are my thoughts on the topic.

It all started when Brian posted on twitter that he was testing RemoteFX with Gabe and I replied saying they should test it with loss. He replied pretty much implying ‘are you nuts? MS is saying RemoteFX is LAN only’ to what I replied ‘WAN is the new LAN so you should test it with loss’.

The reason I mentioned RemoteFX is simple and we must go back a couple years (maybe a decade) to understand what I mean and why I do think ‘LAN only protocols for VDI are Dead on Arrival’.

If you remember (and I clear remember this, back in 2003/2004 when I was working in Japan, accessing my machine over a dial-up connection) years ago all many people had was a dial-up connection to the internet. Things were ‘slow’ at the time and everyone wished they had a much faster connection one day. The idea of having a 1MBit connection, only for you and at home or in a hotel was simply a dream. Everyone though when that day came, all our needs would be solved.

So fast forward a couple years and now, if you are cheap, you are probably using some ‘high-speed lite’ plan from your ISP that is almost certain, at least 1Mbit down/256kbps up. Considering all you had years ago, this should be great, more than you need.

As we both know, this is not the case. Your 1MBit connection is slow. Freaking slow. How come? Well as we can easily see, with more available bandwidth comes all sorts of new technologies like movie streaming, P2P file sharing, rich multimedia experience (from websites, from VDI hosted desktops and so on), etc. The list goes on.

That shows us clearly, no matter how much bandwidth you get in the next couple years, technology will find a way to use it. Either because you will be downloading BluRay2 movies (at 500GB each) or because you need your USB 4.0 WebCam running at 3840×2160 resolution, when connected to your XenDesktop 7.0 hosted desktop (running Windows 9 with 64GB RAM and 32 vCPUs – note it will still boot Office 2015 as fast as a Windows 98 with Office 97 – see Claudio’s Law).

And as we get more and more connected I can only see ‘remote’ workers growing. People that want to work from home, from anywhere and also companies that will start to reduce their office space (that is costly if you do not realize) by giving users what they need at home or anywhere they decide to work from.

That leads us to what I posted on Twitter. The WAN will become the new LAN.

If that is the case if more and more work is shifted to the outside (your home, your cottage, a hotel, etc) are LAN only protocols for SBC/VDI dead?

I do understand that as of today the ratio of users working in a LAN connected desktop and on a WAN connected one is probably 20:1 if not more. But again, is this what the future holds? Will we ever see a shift on this that may bring this down to 2:1 maybe? And if that happens, what future a LAN only protocol has? Type-1/Type-2 hypervisor solutions may alleviate this but again, there may be cases where I do want to work ‘connected’ to my hosted desktop and not from a locally cached copy (i.e. what if I can assign 64GB RAM/32vCPUs to my hosted one instead of using 8GB/2vCPUs for my locally cached one? It will for sure be MUCH faster for several tasks and a reason for me not to use the cached one).

My take on this is, for now, RemoteFX and any other LAN only solutions will do it and will of course help the adoption of a VDI model on the LAN. But as we shift towards an always connected model, if anyone tries to sell their stuff as ‘Good on LAN only’, that will become an issue.

So Microsoft and Citrix, make sure you keep in mind WAN is the new LAN and that whatever crap you develop or acquire in the future has a future on the WAN.

WAN is king.

CR

1,700 total views, no views today

PCoIP performance over lossy networks.

This week as you guys know, I spent quite a lot of time at Citrix Synergy 2010 in San Francisco and during that time we were able to extensively test how all major remote display protocols work over the real world WAN and in a certain way, what we saw simply validated what I was expecting to see.

First of all, let me define real world WAN and explain how we actually know what this is. If you are not familiar with our technology, IPQ, it is a packet loss reduction mechanism that works between two end points. As we adapt according to the network conditions, we must know at any given time what these are. That is the reason why our endpoints exchange a beacon at all times. This gives us over 25 stats that we use to determine the most effective way to deal with packet loss in real time.

Of course all that information gets stored and we can plot what we are seeing right within our web interface. And guess what? During this week we spent in San Francisco, doing our demos from a hotel room using the provided internet connection – the exact same one all of you had in your rooms – we have seen packet loss at all times. How much? From 1% all the way to 15% (burst loss). The bottom line here is simple: loss is guaranteed out there and it is MUCH higher than the 0.5% loss that Brian and Gabe used on their WAN simulator during the VDI Geek Week shootout. That is the reason why ICA, RDP and PCoIP performed relatively well on their ‘WAN’. In the real world, with unpredictable conditions, performance is not really like that. I am not saying that loss will be high and will be there at all times. I am just saying loss will get you several times during the day. When and how much that will be no one knows. But it will be there. For sure.

So back to the topic, how well does PCoIP perform over the real world WAN? Not that well as expected. And here is the living proof of that. Notice how much better PCoIP gets when IPQ is brought to the picture. It gets almost as good as ICA (in case you did not see our tests with ICA, go here).

No matter what VMWare and Teradici tells you, TCP with its retransmission techniques, in this particular type of connection (PPTP VPN), DOES perform much, MUCH better than PCoIP. Just watch the two videos for yourself (the PCoIP is also available in high definition 720p). At 3% loss ICA simply smokes PCoIP (that without our technology is virtually unusable – again, over PPTP. LT2P may change things, making PCoIP closer to ICA over the WAN). The game changes completely when IPQ is on. ICA improves for sure (again, watch the videos) but PCoIP at that point really shines. The improvement is brutal, huge. At the end we turn something that is really unusable over the lossy WAN into something people can actually use. It is that much of an improvement.

In case you want the direct YouTube links, here you have them:

ICA: http://www.youtube.com/watch?v=Yw5lBk-bdv8
PCoIP: http://www.youtube.com/watch?v=AXpbawlg90Y

The bottom line is simple. All current implementations for remote display technologies do suffer over the real world WAN. The idea of our technology is to work as ‘Network Insurance’ for your connection. If the conditions are good we know that thanks to our beacon exchange and at that stage we simply turn ourselves off. But when loss comes, we are there to protect you. Exactly like your insurance company. You have it and hope not to use it. Ever. But when you need it, you know it is there and that you can count on it.

That is what we are. Network Insurance for you and your users.

CR

3,244 total views, 2 views today

ICA behavior on lossy networks.

I guess a picture is worth a thousand words. So what about a video?

Yesterday here at Citrix Synergy 2010 I had the time to record a quick video that shows how ICA, normally a very robust protocol for the WAN, suffers from packet loss. Before you go ahead and say the conditions of the test are not really ‘real world’ all I can tell you (and I can show it in person if you want) is the loss I have seen yesterday over the connection provided by the Marriot Hotel in San Francisco spiked during certain moments to more than 15-20%. So on the real world you will face packet loss at one degree or another. Guaranteed it WILL be there.

This quick test (runs for 6 minutes) shows a XenApp 6 server running on Windows Server 2008 R2 with no load whatsoever. We injected a 3% loss but again, were able to see huge spikes on it (remember, our solution, hardware or software based, sends a beacon between both ends all the time to determine how network conditions are at any given time and adjusts how mildly/heavily we do our magic and with all this data we can plot what is going on over the link in real time).

The results? Well see for yourself. My take on this is XenDesktop/XenApp do suffer. Period. In certain cases your users would experience serious lags when typing, very choppy video/audio and so on. Unusable? I would not go that far. Fixable? Yes as the video clearly shows. And also keep in mind this was all done over a hotel internet connection (the type you get on your room) in a conference where probably every single person IS using the hotel link AND this was done on a XenApp 6 box running in Ottawa, Canada, a couple miles from San Francisco (probably around 3,000 miles).

If you want to understand how we do this (remember, we are a layer 2 solution so we fix ICA, RDP, PCoIP, etc – we do not care what you run; we fix it) feel free to stop me at Citrix Synergy for a chat or just follow me on Twitter (crod).

Bottom line: even though these protocols do have their mechanisms to cope with packet loss, ICA, the king of the kings in the VDI world IMHO, does suffer. If it does, I can only imagine PCoIP will suffer even more (and RDP too). Oh we have tested them.

Yes, they suck.

2,034 total views, no views today

Is VMWare View really easier?

Sorry Brian and Gabe but I must disagree about this. In the past two weeks we have been working on getting our demo environment up and running (so you can all see what exactly packet loss does with VDI/SBC) and we will be showing it live at Synergy to interested parties, in our suite at the Marriot.

As I still think overcommitment when properly used is a powerful feature to have and the fact that XenServer 5.6 is in beta, we setup our virtualization backend with VMWare ESXi and I still must say I really like it. I did not have much exposure to XenServer and Hyper-V so I cannot comment how they compare to ESXi but again, for what we need ESXi does the job beautifully and it is indeed pretty straight forward to get going.

So back to the VDI part, we asked one of our technical guys to setup XenDesktop and VMWare View, based on the ‘instructions’ I gave him (basically download the ISOs and follow the guides) and off he went.

As I do have my own XenDesktop environment at home (on ESXi as you remember from the saga I reported here) I was very familiar with it and do remember having no issues installing it and more than that, being extremely easy and simple to do. Keep in mind this is the free 10-user version so there is no Provisioning Server, etc.

Today I sat down with Matt (our techie) to go through the environment and to wrap it up by setting up a 2008 R2 RDS SH with XenApp 6. As we have been exchanging emails to get VMWare View going, I knew it was not being that breeze Gabe/Brian mentioned. So after checking what we had to go through, I have no idea why they mentioned VMWare View being ‘easier’ to setup. IMHO (and in Matt’s opinion as well) XenDesktop 4 is much, we mean, MUCH easier to setup. VMWare View is confusing to say the least, even for someone very used to a single ESXi server environment.

How are they affected by packet loss? Well that is what I am working on right now. The same way Ruben wrote the ‘VDI Storage Deep Impact’  I am working on getting the real facts on how latency/loss affect ICA, RDP and PCoIP. No marketing BS, just straight facts. As soon as the whitepaper is ready it will be posted here for sure.

Resuming this post: we found XenDesktop 4 much easier to setup for SMALL ENVIRONMENTS (what means Provisioning Services/View Composer are NOT in use, that you do not have a VMWare Guru/Employee sitting on your side, etc). Does it mean Brian/Gabe are wrong and we are right? No. It means you should take any comments, from anyone, Brian and myself included, with a grain of salt. Do not trust us gentlemen. 🙂 Go try them for yourself and let us know what you think.

As you can see I do like VMWare and do think they have a great virtualization platform (main reason to have ESX/ESXi all over the companies I manage/own/work for) but View is far from being a breeze to setup.

CR

49,755 total views, no views today

XenDesktop 4 on ESXi. Final chapter on this saga.

As you guys do remember, a while ago I posted about my experience on trying to get Citrix XenDesktop 4 working on my VMWare ESXi environment. You can read the post that started this soap opera here and an update on what happened next here.

First of all I must say that post started quite a discussion and more than that, it is the #1 search that is now leading readers to my blog, especially because most people think that as a Citrix CTP, I am not supposed to post things like that.

It is quite the opposite actually. Of course when I started that thread, Citrix immediately contacted me but not really because I did it. They contacted me to understand how they failed on not supporting me or not getting me answers for all the issues I had at the time and promptly started working with me very closely to get that environment up and running. And in the end I must say, it DOES work and DOES work as expected.

The outcome from this post that started it all is this: Citrix today announced they are OFFICIALLYsupporting VMWare ESXi as a host solution for their Citrix XenDesktop product! You can read their official statement here: CTX124952.

Of course there are certain features regarding the integration with the Hypervisor that will not work (Pool and Power Management ones). For a small environment, with let’s say 5 to 10 users in a company, this is in my opinion, great news. It is an easy way to dip your toes into the VDI world at pretty much no cost. Build a small server running ESXi (free, or use the paid version that is on sale at $495 until June, 15th I think) and create your XP/Windows 7 VMs on it for your users. Again, for a small shop I do see this as a great solution. Pretty much all you need is there, including a Web Interface, Gateway, etc. Unbelievable value at no cost. And if down the road you do feel you need the features available on the paid versions, it is not hard at all to upgrade.

And technically, what fixed the issues I was having is all that is described on the second post on this thread ANDa ‘power’ setting on Windows 7. Apparently by default it will go into ‘Standby’ after a couple minutes. Once I changed all power settings not to turn anything off at all ever, the problem was resolved. I can now consistently logon to the Web Interface and get to my XenDesktop 4 Desktop.

So resuming: XenDesktop 4 DOESwork with VMWare ESXi ANDit is now fully supported by Citrix, thanks to your friendly neighbor CTP here. 🙂

By the way, it works beautifully.

CR

2,010 total views, no views today