|
Hi Mia,
> What's authorization cache?
The FCS caches the authorized access to a drawer and refreshes that
information once every while. Basically who has which access to a
specific drawer. It's held in cache for performance reasons.
> Is there a document that explains threads, connections, tasks, etc.
Please refer to notes 203, 205 and 304 in the conference. Also refer to
note 769.1 (Terry Porter) for more information on a particular document
written on the FCS.
Ciao,
Simon
|
|
re .0
Mia
as .1 said the authorization time out is the drawer authorization.
For performance reasons, when the FCS opens a drawer, it caches the
access info on teh drawer, info like the ACL, prot mask etc.
Every once in a while this info is flushed and re-read. The ensures
that the FCS is not granting access based on stale access data.
When you change the info the FCS that changes it will refresh its cache
then so that FCS will have the new info immediately. The potential
problem that the authorization TMO gets around is that if you hve
multiple servers running on a cluster, the access info is cluster wide.
If srvA does the access modify, then if srvB also has teh drawer open
it will not pick up the new info until it refreshes its cache.
THREADS -- The FCS is multi-threaded. You can almost think of this
as being parallel processing. Each request made of the FCS
is given its own thread of execution. What this means is
that the FCS can handle multiple tasks at the same time.
Tasks do not block waiting for the one infront of them
to complete. It gives the appearence of multi-processing
or parallel processing. Clear enough? (if not ask more)
CONNECTIONS -- This is the count of client connections made to the
server. Every time an IOS user opens a non-default drawer,
or does a cross drawer operation a connection is made to
the FCS. This will also include ALL TeamLinks connects.
TASKS -- Each request made to the FCS is called a task.
Any more??
--bob
|