Hi, At this years (2013) Innovate conference I spent quite some time with Jane Balin discussing our requirement for a distributed license server model to support the need to operate over slow network links and to also guarantee licenses to certain projects. I have not heard any more about this so an update would be appreciated.
Ideally it would work as follows.
1 Central server with all tokens/floats located on it. At each site a cache license server which can be configured to hold a min number of tokens locally at that site (so the cache servers would check them out of the main server and hold them locally.. requesting clients would then check them out of the cache server). As the cache servers local pool of licenses dropped below the minimum threshold it would pull more out of the main server... likewise once a license was returned to the cache server pool it would also check back in licenses to the main server.
These cache servers will reduce any latency issues which we suffer on a number of sites and keep the apps performing at the best level. Licenses could also be installed on these cache server such that if a license can only be used by one project (for contractual reasons) then it could be just on the one special cache server and the cache server would pull all extra licenses from the main server as required.
Enabling such a topology would improve performance and robustness, allow WAN usage of tokens/flats, allow true global licensing and allow companies to increase their licenses utilisations and decrease maintenance of manually having to shift licenses around using the License Key Center.
This topology would also need to be supported by the License Reporting and Analysis tool although I think it would already more or less support it in its current state.