Archive for February, 2008

Anyone how has even done a little bit of Microsoft Terminal Services or Citrix WhatEverWeAreCallingItThisYear 4.5 will know that THE biggest pain point in a thin-client environment is printing. Sure we have come along way from Windows 9x drivers getting “accidentally” installed to Window NT4.x TSE and causing a BSOD. But a lot of back slapping from MS and Citrix is not the order of the day. Printing is still a problem. What exactly are the issues? Well, for the uninitiated thin-client virgin (is that you? Keep your head down, nobody will notice) it boils down to this. Printer drivers suck. I speak from a position of knowing some guys at HP who will involved in printer driver management/creation who almost had nervous breakdowns (it’s a long story) because of damn things.

Anyway, rant apart its stuff like massive ballooning of print job size – from a file that 10MB you can have print job that’s 200MB. You try putting that down a 128kps/256kps/512kps pipe. It’s call network constipation. You gonna need a major suppository after trying this – and will be looking for solution pretty quick. Secondly, solutions that have attempted to solve this inherent logical problem – the Virtual/Citrix desktop is in the Corp Datacenter is in Palo Alto but the user is Papa New Guinea have invariable compromised functionality in order to reduce job size. Whether its Citrix Universal Printer Driver (oxy-moron/contradiction in terms anyone?) to UniPrint’s PDF solution.  By no-means the new-kid-on-the block – thinPrint (founded circa 1999) is a solution which I’ve never tested/used/played with. So in the spirit of curiosity and also this VDM project I mentioned earlier I attended this session. Other issues you find with printing aside from obese print jobs – is the mapping process. How do you map network printers and client printers in the RDP session, to the actual destination device.  Sure all the major vendors have a technology that simplifies the problem. They call it by number names – but the main word is “Automatic”. You see this is a sunny day process – sometimes it works, and yes you guessed sometimes it doesn’t. Anyone who’s had fun with Wyse terminals on the CE platform will know what I’m talking about here. But hey, never mind – just log out and login again.

This is the just kind of nitty-gritty thorn-in-the-side, PTA issue that curiously does not get address in C-class Keynote session. The inconvenient truth as some guy in the US put it. You might not know this that this apparently a very strong relationship between Citrix and VMware. From what I’ve heard there is an awful lot of bodies at VMware who are former Citrix personnel.  So you would think this one that VDM would have well taped. However, it does seem that VMware is leaving this up to the mature 3rd party print solution market. Make sense. That would be expensive, plus it gives you someone to blame it all on when it doesn’t work.

Themed – water, ice, fire and earth. The VMworld party was spread across two big floors round the back of the convention center. The evening parties at the EMEA events have never been as good as the US events, and the last two TSX’s have for me always been in the shadow of the Paris TSX. You see when it’s a choice between free rides on rollacoasters at DisneyLand Paris or a man who covers himself in baby oil and does something odd with a weather balloon – I know which one I would prefer! [Please don't email me ironically asking which one it would be!]

 I spent most of the evening with a customer/pal of mine Sandy, who works for Ballie Gifford. Ballie Gifford a big finance company up in Edinburgh and were one of the first customers I taught Vi3 on when it was release in September 2006. Gee, that date sounds like such a long time ago already. Scary. So me and Sandy had our food and beers – and sought out entertainment.  We did a spot of LaserQuest – which involves a lot of darkness – and firing Laser guns at each other – a chance to be RobCop for 5 mins. Next we bumped into Maxim from Veeam who was accompanied by two lovely ladies of Veeam. The ladies work for Veeam and taking time out from the University studies in Economics and PR/Marketing to get some real world experience before writing up the whole thing for thesis.  Sandy and I took five minutes to direct them to where the food was laid out, out and to temporarily bask in the glory/glow of the girls wondering how we had got so lucky to be in their presence! More importantly making my colleagues from the London VMware User Group look at me and Sandy all-a-gog!

Speaking off lovely ladies, we went back up stairs to watch the out-door pyrotechnics of a blonde and brunette duo. These two did a ten minute performance with a glorious 40′s noir backing track. All I can say is – I now have reason to go back and explain to my girlfriend why we do indeed need an mini-angle grinder at Palias RTFM. Say no more!

I’m sure these performances will wind-up on youtube shortly. I’ll see if they surface before I post this entry.

I attended this session mainly to meet two guys Chris Wolf (Burton Group) and Mike Dipetrillo (VMware). I’ve had email exchanges with Mike concerning various issue of external support, licensing and so on with other vendors. I’m hoping Mike will run an eye across a presentation I will be delivering for a UG in Charlotte, Carolina in April. I see Chris’s name all over the net presenting topics related to virtualization – I figured he was a personality on the virtualization circuit with whom I should be acquainted. Anyway, Mike did the scene setting but it was basically Chris is gig telling us all what’s nasty and bad about licensing in the world of virtualization. It certainly was an eye-opener for me – as someone who doesn’t really get that involved with the purchase and licensing of software. It was precisely because of this lack of exposure that I attended this session, so I could soak up the benefit of Chris’s impressive vendor-wide knowledge. I don’t know how he does it – as he said himself if you have insomnia problems, a 60-page licensing summary from a vendor is a great cure. 

What was interesting about this session is how 4-5 years on from when I first looked at VMware the software industry hasn’t really got to grips with this darn thing called virtualization especially our friends Microsoft. Expect a complete route and branch reshaping of the licensing policies from M$ with the release of Hyper-V. Not because M$ are going to make substantial cheaper to run their software products to run in a M$ VM, but to make it substantially simpler [well, we live in hope]. These changes should make it better for all customers regardless of what vPlatform you choose to adopt in your datacenter.  What was missing in this session was an assessment of VMware’s own per-socket licensing model, and how sockets based model has specific implications for ESX customers because of the VMotion CPU requirements. I’ve heard of some crazy stuff in the field because of this – like people removing factory fitted CPUs (because they don’t need the horse power right now) so they don’t have to license a socket they aren’t using. BUT, at the same keeping hold of them when they need them – for fear of introducing a CPU incompatibility by a later upgrade. Sounds a little crazy, and depending on your HW procurement model it shouldn’t be necessary.

Anyway, for products like SQL and Exchange MS 90-day movement prohibition with licensing is a number of things. Firstly, directly in conflict with principle of liberating yourself from your hardware with features like VMotion and Citrix Migrate (or even M$ own “Quick” migrate). Secondly, expensive if you foolish follow the letter of the law/EULA and lastly unenforceable.  Just in the legal world an unenforceable EULA is just an extra piece of unneeded legislation that benefits no-one. At this point Chris conceded that if he wasn’t careful he could be accused of promoting customers lying to Microsoft. I piped up to say that in the meantime – until M$ tidy up the mess they have created for themselves and their customers, that lying was both moral and legally indefensible – but being “economical with the truth” (a very British phrase) shouldn’t be a problem.

I woke up on the last day quite tired. Met Sandy down in the lobby (we turned out to be staying in the same hotel) and got a taxi with all gear to the convention center. I head off to the very long titled breakout session called [deep breath]  Virtualizing I/O Intensive Application using VMWare: A Microsoft SQL Server Case Study.

I know practically zip about SQL, and you know I’m more than happy for that to remain that way. But what I am interested in is performance. One of THE most common question I get from my students is should I virtualize [insert application server name]. You know like should I virtualize citrix, sql, oracle, sap…. and so on and on. This session was presented by a chap from Brocade – and I was looking for some real stats independent from VMware that would give me some ammo in the classroom – to back up my answer of YES, unless its some huge 8-way or 16-core box which is using 32GB of RAM. Brocades tests were relatively modest – singe CPU VMs with 2GB of RAM. That’s a profile which doesn’t really reflect the common physical hardware assigned to SQL on a PM.  None the less the result were very good – and Brocade are promising more research. Ironically, they discovered their SAN configuration was the cause of an unexpected bottleneck. Additionally, some (but not all) of their CPU stats were skewed. I asked them whether this was because they were running performance collections inside the VM (a common novice like mistake). It turned out that there is some kind of flaw/error/bug/deficiency in the collecting performance metrics from VMware. That’s a little worrying – apparently VMware aware of this and there will be a patch. I should have really got hold of this guys contact details so I could have pursued this further.

I was very kindly invited to an evening meal with Veeam.com. If you don’t know Veeam, where have you beean? They’re the guys behind the Reporter, Configurator and that FastSCP Tool for ESX. I recently did a webcast for the TechTarget.com guys on the subject of storage – and the Veeam people were very kind to sponsor the entire thing. Whitepaper to follow. I first became acquainted with Veeam last year, and was invited to their dinner event at VMworld 2007. So it was chance to re-enforce that relationship. They are a really friendly bunch of guys. I caught up with Doug from Veeam who was the moderator/interview for the webcast. So it was a real name to face situation. I also got a chance to quiz Doug and Carrie who work Veeam about their backgrounds – and how they found themselves to be work Veeam. As ever in our industry Veeam’s almost like a little family with everyone having worked for previous companies in the past – liking and respecting each other – and deciding “hey, let’s make a company”.  With the history of Russia being what it is – it’s this generation of people who are really at the forefront of trying to foster, created, generate and protect a true entrepreneurial spirit out there. I’m no hard-nosed “Thatcherite” at heart but I have an awful lot of respect for people who prepared to put their heart ‘n’ soul into something they believe in.

 There was a couple people there I wasn’t expecting like Tarry and some of the instructors who work for resellers in EMEA. It just goes to show this whole virtualization thing is still a very tight community of enthusiastic individuals (like me!) all the way from USA, to the UK, and to Russia.  

I hope you catching a theme with my breakout sessions. They are all about new products. I’m not attending sessions which about tuning, optimising or troubleshooting existing known technology from VMware. Done that at Nice, Paris, LA and San Fran. When I look at new product whether it’s from VMware or another vendor – I’m doing it from a purely selfish perspective – I’m looking if there’s a financial opportunity for yours truly. It may not surprise you that I don’t RTFM Education LTD as not-for-profit entity. Until the guys from forums get together to pay my mortgage that’s like to continue to be the case!

So. What is Mr RTFM looking for in a new product. Well, my strategy for success (sounds like the title of some horrid self-improvement book) is Skill, Hardware, Customer.  Firstly, do I have the skills either personal or technical to get this product under my belt – or more specifically what amount of investment of my time to get those skills would have to invest to get a return. Secondly, hardware – does the technology have specific hardware requirements – and depending on heavy they are – can I buy that, or can I emulate it. A good example of this is SRM requirement for replication. I have a SAN at home. Good. But not two. Bad. Together with that – I’ve got no replication technology. I’m hoping I will be able to put together such a system using virtual appliances from Lefthand Networks or NetApp. Lastly, Customer – does the product have a clear, recognisable usage case such that a customer can see right away “what the product is for”. If customer can see this – I needn’t waste my and the customers very precious time in selling it to them. You see I’m a rubbish sales man. The best products are ones that sell themselves. VMware sells itself. Citrix sells itself. The less said about Microsoft the better.

Anyway, joking apart – I wanted to attend the ThinInstall presentation because I know little about it. It’s a recent acquisition from VMware which will slot just dandy into their VDI solution – VDM. You see alongside SRM, VDM is going to be my next project. But I won’t be just looking at VDM in isolation. My guide to VDM will have to be multvendor because right now VDM solutions are blended ones that involve a range of technologies. So this guide will probably include some DeDupe and Cloning from a storage vendor such as NetApp – then the VDM piece with all the features configured and setup – ending with ThinInstall to deliver the applications.  Lastly, I want to include some kind of review of 3^rd party printing solutions such as UniPrint (which I’ve not looked at in years!) and ThinPrint. There’s little point in looking at VDM on its own.

[Watch out ThinPrint, VMware just bought ThinInstall. VMware's got a thing about thin... JOKE!]

The ThinInstall demo was very interesting and trip down the recent memory lane of being a Citrix guy packaging up MSI’s and Citrix Application Manager apps for mass deployment  not to workstations but Citrix WhatEverWeAreCallingItThisYear 4.5. Seems to be that ThinInstall wins because it is a.) very easy to use and b.) agentless. Your left with a flat file at the end of process. I love that. When I find a free application on the world-wide-interweb and it doesn’t have an installer I love that. If I don’t like the application I don’t have to worry about trying out in a non-persistent VM or de-installing it from my physical laptop. The nice thing about this agentless approach in comparison to SoftGrid is the paradigm of having to install one piece of software to not have to have to install any other software is a bit, erm – weird.

After my chat with Tim, I bumped into Tarry Singh, and was roped into an unplanned and unscripted discussion with myself, David Marshall, Blogger, and Alex Pelster, Sr. Consultant, AtosOrigin about virtualization issues generally. I won’t repeat what we rifted on at the time. I’ll leave that Mr RTFM – Mike TV…

http://virtualization.com/video-audio-vodcast-vlog/2008/02/27/video-discussion-with-four-virtualization-analysts/