Guest post on the recent announcement of more Autodesk Citrix Ready applications by a longtime Autodesk customer and experienced programmer of product install, deployments, and customization
<drum roll> David Stein
At the risk of sounding like a flip-flopper, I need to shake my dented frying pan brain and flip the omelet over for a minute…
It started with Autodesk's press announcement. That led to Shaan's blog post. That led to a colleague thumping me in the head to wake up and read Shaan's post (because I forgot to check up on my RSS feeds for a whole day). That led to me picking my jaw up off the floor, brushing off the dirt and cat hair and put it back on. That led to me posting about the significance of the announcement. I only posted a brief opinion on it though. Rather than drill into a monotonous monologue, which I'm usually known for, I thought that it might be best to approach the significance using a familiar FAQ format.
What was the announcement really about?
After years of being prodded (by a small percentage of customers, mostly large-scale environment customers), Autodesk decided to put their support behind the use of a centralized installation of nine Autodesk 2012 products on a Citrix server using XenApp. This allows clients to run AutoCAD 2012 from a remote/centralized server rather than having to install it locally.
Why is this a big deal?
Put simply:
- Would you rather install 20,000 AutoCAD 2012 clients, or one server installation?
- Would you rather deploy a service pack to 20,000 clients, or install it on one server?
- Would you rather upgrade 20,000 clients, or upgrade one server?
- Would you rather buy RAM upgrades for 20,000 computers, or have them run the app remotely?
Who stands to benefit most from this?
Most centralized server-hosted application services are found in large corporate data center environments. Small and Medium sized businesses may not find this to be a practical solution due to costs required to adequately implement a Citrix server environment, and staff and support it as well. There is no hard-set number of clients or licenses that define what a "large" environment is. In my (humblest) opinion, when you are installing hundreds of licenses (or more) you may want to look into this option.
If you already have a Citrix implementation, you likely have a “data center”. If this is true, and if you deploy and maintain a large number of Autodesk 2012 products (or soon will), this is like eggs and bacon. The only real difference now is that you’ve been given the green light to put them in the same frying pan and make something tasty.
What impact does this have on traditional client deployment environments?
Short answer? None. You can continue on with FLEXlm network-licensed deployments, or standalone license installations just fine. A Citrix implementation of Autodesk 2012 products does require a change to the FLEXlm license, so contact your Autodesk reseller for details.
What impact does this have on standalone deployments?
If you have a lot of standalone licenses to contend with, and you already have a Citrix implementation, this might be a good time to consider moving them into a hosted configuration. That will depend on their connectivity obviously. If the current standalone license users work offsite without VPN capability, you might want to continue as-is.
What can I expect from performance and user experience with a Citrix/XenApp approach?
Performance and User Experience are almost always a functional result of how much money is invested. How you invest in your servers, SAN, network, routers and switches, and the logical configurations will all directly impact what the user experiences on the end of it all. If build a shoe-string budget implementation you should probably expect a shoe-string budget result.
For large corporate environments, where this will almost always involve multiple departments, it is imperative that everyone involved has the same goals in mind as far as providing an optimal result for the users. If one of the team members doesn't take the project serious - it is going to fail. Get buy-in from the highest ranking suits in your company, especially the ones with the most fist-pounding temperament.
What about the risks?
Risks are usually mitigated with money and effort. What are the biggest risks?
- Single point of failure
- One-to-many dependency
- Connectivity impacts
The first risk can be mitigated with server farms, clustering, load balancing and tiered services. Much of that will depend on where the users are located. If all of them are in one building, the solution is usually not very complicated. If they’re spread around the Earth, it’s likely going to be more complicated.
The second risk can be mitigated with tiered services. Spreading the server farm to spread the load and the connection points, thereby reducing the distance or latency between various groups of users and the nearest Citrix connection.
The third risk is something that can (or should) be addressed by your network engineers. They will know the topology, performance and capability aspects of all the links between data centers and user groups.
For answers to your concerns, contact a Citrix reseller. They will work closely with your network engineers and your server managers, in order to design a solution for your needs.
Where will I need to spend the most money?
Review the Autodesk Hardware and Network Recommendations for Citrix XenApp guide at http://autode.sk/g2tl3i. Using this as your guide, and consulting your Citrix and Microsoft support representatives, you should be able to design the ideal configuration and then determine the gaps between where you are now, and where you want to be. Closing that gap is where the costs are revealed.
How do I know if this is a good fit for my company?
How do you know if any new technology is a good fit? You assess your current situation, what problems it creates and at what costs (get down to exact dollar figures). Then you work out the solution plan and assess what it would incur in terms of problems and costs (dollar figures again). Then assess what it will cost to implement the new solution. Compare the before and after costs and BINGO! You have a clear idea of what to do. That is what you staple together and put in front of the executives in order to sell your idea.
How do I identify the costs?
You need to combine all the costs together, but they most often break down into the following aspects:
- Hardware purchases
- Hardware Support costs (vendor or reseller)
- Software and Licensing purchases
- Software Support costs (vendor or reseller)
- Training of IT staff (classroom, books, web, etc.)
- Lab testing (hardware, software, networking, facilities, time and labor)
- Network Services (routers, switches, wiring, ISP fees, support costs, etc.)
Data Center Stuff
Citrix server farms can be virtualized. Imagine that. You can virtually virtualize a virtual solution. So, if you already have a robust investment in a server virtualization environment (VMware, Hyper-V, Xen, etc.) you can leverage that to avoid purchasing additional hardware. This is also dependent upon what SAN resources you have available. After all, when you’re talking Citrix, you’re often talking data center, and data centers often involve virtual server hosting and SANs. Consult your Citrix reseller for help with assessing your readiness.
Disclaimers
I am not a Citrix consultant. I do not work for a Citrix consultant, nor do I have any family, friends or neighbors that work for Citrix, a Citrix reseller, or a Citrix consulting group. I work for a Microsoft/VMware/Cisco consulting firm, and bump into Citrix admins at client sites and have meaningless discussions during breaks about stupid meaningless topics. My direct knowledge of Citrix implementation is very limited, but I do understand the nutshell aspects enough to act like I know what I’m talking about when I’m holding a mixed drink at parties and cookouts. Don’t ask me what I said the next day though.
David Stein
Thank you David for your thoughts, suggestions, and humor.