βTechnology Consulting Group Logica Australia

This time it all started following a meeting with Riverbed sales reps in Sydney and Steve Orkbi throwing at us a couple of links about the recent move by Mirvac. Here is the chat summary:

Steve Orkibi: You may have seen these already, but there have been a couple of interesting articles recently relating to Mirvac’s decision to move from a Citrix-based solution to a Riverbed WDS solution for their regional sites.

Here is the original article: http://www.zdnet.com.au/news/communications/soa/AU-100K-switch-to-Riverbed-connects-Mirvac-staff/0,130061791,339284174,00.htm

An the follow-up article by Citrix: http://www.zdnet.com.au/news/software/soa/Citrix-hits-back-at-unworkable-CAD-accusation/0,130061733,339284298,00.htm

Angela Eksteen: Riverbed would be my choice for WAN compression very good for cross Data Centre connectivity I know they have small units suitable for small offices not sure how cost effective it is though although given the price of Citrix now I’d probably even go TS over Citrix (once Longhorn is released).

Kirk Downey: Perhaps for DC to DC, but Packeteer also does Wan Compression. For branch offices, though, Packeteer’s black box runs W2003 native, so you can not only push file storage back to the DC, you can also replace local Domain controllers, exchange nodes, file and print, etc with the same black box. For branch office apps, you can get the same product as software (opposed to black box) meaning you can do all of the above on a single site server, but also run virtual sessions on the same box to run business apps. Net result is a consolidation to a single component for branch and/or remote offices. Couple this with some decent bare metal deployment technology such as altiris or SMS and you can deploy and/or recover remotely. This is the interesting thing I was referring to. I used packeteer in the UK for large regional offices and call centres <…>

Angela Eksteen: They all have their uses I expect. Riverbed offer similar sort of functionality as packeteer and they have small units designed for branch and home offices.  <…> I suppose it comes down to what best fits what the customer wish to do.

Kelley Johnston: Citrix used to be sold as a way to preserve investment in desktop infrastructure.  Considering the drop in desktop PC prices over the last few years and their much larger capacity since Citrix hit its peak and the rise in browser-based applications, is Citrix going to have to re-invent itself in order to survive?  I can’t see that original argument holding a lot of weight any more.  Unless that question is answered I wouldn’t be comfortable making any long term investments in the Citrix camp.  Not because their product wouldn’t fit the bill, but because the above issue might compromise their stability as a company.

Jonathan Hambrook: I would agree 100% with Kelley, Citrix is holding onto market share, the new version have little new technology and from person experience the only cost benefits I have seen are:

1. When your users have 1 or 2 applications they use to perform their job. If this is the case the amount saved on software and hardware would be a benefit. However if they require specific applications and one off’s then this is where the expense can be an equivalent to that of a desktop/laptop with out their benefits.

2. When you need to deliver a standard set of applications over a small pipe. Great for branch offices etc.

Having said that their product is good, the last environment I was exposed to it had 1500 desktop users running a Citrix desktop off WISE terminals. However this was in insurance and all those customers only used Email, Word (Letters/Faxes) Web Browser and Insurance Policy Application. Simple and cheap for the masses, they never worked from home or needed to store data locally.

John Rose: Besides potentially reducing hardware costs or minimising desktop refresh the biggest selling point for Citrix is reduction in operational costs specifically in the area of SOE management. Instead of deploying and managing multiple end points and dealing with subsequent sociability issues a central philosophy was supposed to reduce effort but in my experience this wasn’t always the case. Administrators seemed to have lots of problems coming to terms with the rigours involved with managing a multi-user environment. Overtime the number of users per server would reduce and cost savings and stability evaporate. Many IT decision makers were burned when business case and justifications fizzled!

Things have improved markedly with introduction of isolation tools and experience however this has also led to the biggest shift from traditional Citrix. Application sociability absorbs much of the effort that goes into the development and change management of a desktop or Citrix SOE. Many third-parties came into this space attempting to fulfil a need. Initially isolation tools allowed previously unsociable applications to cohabitate reducing total number of servers required in a farm.

Softricity went further and actually managed to virtualise applications so they could run in their own separate space. With the likelihood for application conflicts removed a major reduction in the sociability testing effort associated with application packaging/deployment resulted. Better yet this type of application virtualisation allows a client to continue running an application even if the publishing server is not available. This was another problem for traditional Citrix, all eggs in one basket. Cheap hardware coupled with excellent SOE development methodologies such as Business Desktop Deployment and the desktop SOE became attractive when compared to traditional Citrix.

Of course Microsoft purchased Softricity and will rebrand Softgrid Microsoft Application Virtualisation. Incidentally Citrix also has an application virtualisation solution bundled into its suite and considering it already has an established base of users I’d have expected to have heard more about this offering. I suspect that the Citrix brand is too entrenched into the traditional application publishing architecture and the market is being attracted to big blue and especially the pioneer application virtualisation product. MAV is only available to Microsoft EA customers that purchase Desktop Optimisation Pack licenses which off hand are around $15 per seat, not cheap but inexpensive when compared to sociability testing.

Bottom line for me is Citrix is being hurt by market perception and a major reduction in the costs associated with managing desktops and provisioning decent WAN links. I have no idea how I have ended up with this conclusion based on Steve’s original thread, but there you go!