Mark Pottorff wrote:
> It would seem that having the server be able to accept a connection,
> and then direct the client on any preferred backoff would be ideal.
[...]
> establish a simple 1-5 system expressing how busy it is. The client
> then weighs time to deadline against likelihood of future internet
> connection to assess whether it has an upload of sufficient priority
> or not. And if not, it delays starting next upload.
> 
> Really the objective is that all result files across the user base
> get uploaded in priority order. Where priority takes in to account
> bandwidth user and project have available, deadline, file size, etc.
> 
> Same is true for downloads. If server is drowning in requests, you
> want to be servicing the starving clients first, not those with
> full-time networks and fat work caches. And only to the extent that
> they aren't starving anymore, until the contention is resolved.
[...]

Good points there I agree with.

Do not forget that there are still users that are only on intermittent 
56.6kb/s DIAL-UP connections...


A lack of traffic management is indeed a BIG problem.

The software people cannot just shrug it off as a "network problem".

In reality, traffic management is a problem affecting the entire system.


I rather like the idea of clients requesting a time slot from the 
project server for when they can upload their data and at what rate...

Hence have:

client -> server: Request to upload at xxx-kb/s, I'm on yyy link, ccc 
available cache;
server -> client: Granted at zzz time.

Where yyy says how long this link is available for and how frequently it 
is available. (eg: Numbers that indicate dial-up?!)

(Aside: User client-sourced information will avoid loading the central 
db... Or would that be too open to abuse from "modified" clients?)


However, in the immediate scheme of things, I think the dynamic tweaking 
of client upload rate and of the backoff parameters and a slow decay to 
the backoff times are more easily achievable. Dial-up users take their 
chances but with a great disadvantage to all others during overload 
conditions.

For a quick fix, perhaps the project servers should just give an 
immediate NACK to excess connections when the link is more than 80% 
utilised. Tweak that number for maximum throughput during recovery periods.

Regards,
Martin

-- 
--------------------
Martin Lomas
m_boincdev ml1 co uk.ddSPAM.dd
--------------------
_______________________________________________
boinc_dev mailing list
[email protected]
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.

Reply via email to