LSTOWN-L Archives

LISTSERV List Owners' Forum

LSTOWN-L

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Topic: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Francoise Becker <[log in to unmask]>
Tue, 18 Apr 2000 16:34:47 -0400
text/plain (37 lines)
On 18 Apr 00, at 14:19, Pete Weiss wrote:

> Unless you're using the hiperf. version, I believe adding thousands of
> users will be time consuming.
>
> I don't have any stats myself -- just rumors.

See http://www.lsoft.com/products/default.asp?item=listserv_perf

I once did a calculation in response to some poor soul who was
trying to add 60K subscribers to a LISTSERV *Lite* list. Based on
the benchmark figures in the above link, I roughly calculated that it
would take something along the lines of 39 hours to add 60K
subscribers in LITSERV Lite, just under 3 minutes in LISTSERV
Classic, and a few seconds with LISTSERV HPO.

Generally we recommend that Lite be used for lists no bigger than
1000 subscribers (and only a few lists on the server). If a site has
more than 50 lists or over 50K subscribers total, you should
consider HPO, and you should definitely get HPO if you have over
150 lists or over 150K subscribers total. But it's really a "quality of
service" issue. If you don't mind that commands take 2 weeks to
execute, you could probably use Lite for a large list ;-) I have not
tested, nor do I ever intend to test, this statement, though. The
point is that these numbers are just rules of thumb: if you have
Classic and you're happy with the perfomance, don't go out and
pay extra for HPO simply because you match the recommended
numbers. If you have poor performance and you want to improve it,
go ahead and get HPO even if you only have 40 lists or 40K
subscribers. And of course it depends on how the subscribers are
spread out. One list of 50K may warrant HPO but 10 lists of 5K
may not. And there may be other things you can do to improve the
performance besides turning to HPO, such as more memory or a
faster CPU, or experimenting with the FIOC parameters.

Francoise

ATOM RSS1 RSS2