PJ_47272
Nimbostratus
NimbostratusDeployment Design
JP Is there a reference to use for designing support for implementations > 10,000 CCU? i.e. which platforms are required. Also, Citrix seems to measure support in terms of users and not connections. Each user could have 4-5 applications open and depending on what kind of application it is or whether it is xenapp versus xen desktop, the bandwidth consumed could also vary wildly. Multi-stream ICA seems to exacerbate this. Taking these things into account, should we be designing for connections or users with APM and are there numbers to refer to that detail which platforms APM requires to hit these limits? Did some trolling around the site but could not find anything...Thanks in advance.
JP, a couple things here. F5's sizing in terms of XenApp / XenDesktop is predominantly user-centric. Regardless of how many streams, applications or connections a user establishes; they still count as a single concurrent user.
In the example of 10,000 users, this is the maximum capability of a BIG-IP 4200v which supports up to 10,000 concurrent users and 10M concurrent connections. If each user was leveraging multi-stream ICA with 4 channels per application and say they had 5 applications open, that's only 200k active connections out of the 10M capacity; well under the connection management limits.
For deployments larger than 10K users the BIG-IP platforms scale up to support additional user volumes as your environment requires; for example 60K users on a 10200v or up to 200K users on Viprion 4800.
