Upper bounds of TCP & server processes in Unix

Thread View: comp.client-server

3 messages

[T] Back to Threads [V] Tree View [G] Group Articles

3 total messages Started by Eric M. Carroll Wed, 04 Nov 1992 19:58

#35

Author: Eric M. Carroll
Date: Wed, 04 Nov 1992 19:58

33 lines
2247 bytes

An issue that I have seen kicked around briefly in these groups, and
now has arisen locally, is a question about the upper limits in Unix
for supporting large volumes of TCP connections and server processes.
There are many examples, but email is the one we are considering.  The
question is, given a Sparc 2 class machine and SCSI disk drives, how
many connections can it usefully hold open, and how many server
processes can it run simultaneously? Assume no users actually log in.

As a performance analysis problem, this clearly has many components,
including size of the process, duty cycle and traffic activity of the
client/server connection, amount of file activity per sever, and many
other things.  One way to answer the question is simply to try it with
a synthetic load and measure the response curves. Another way is to try
and understand the hard, intrinsic kernel limits, such as the upper
limits on file descriptors, the non-optimal searching algorithms for
TCP control blocks, routing tables and other heavily accessed network
data structures (eg. linear linked lists instead of hashing), process
table size limits, buffer limits, etc, etc, etc.

Has anyone tried to drive this class of machine hard in a
"connection-oriented" client/server relationship? Is there any relavent
performance analysis literature in this area? Or does anyone have any
fundamental insight into this problem based on their knowledge of
Unix?  Any comments on this topic would be welcome. If there are
vendors that have specifically addressed these issues in their kernel
implementation, I would like to hear about it.

 

-- 
Eric Carroll            University of Toronto Computing & Communications
                        Network & Operations Services, Network Development

Re: Upper bounds of TCP & server processes in Unix

#54

Author: Brenda Parsons - x2403
Date: Fri, 06 Nov 1992 20:21

31 lines
1846 bytes

In article <Bx7JGy.6BJ@gpu.utcs.utoronto.ca> eric@cathaus.utcs.utoronto.ca (Eric M. Carroll) writes:
>
>An issue that I have seen kicked around briefly in these groups, and
>now has arisen locally, is a question about the upper limits in Unix
>for supporting large volumes of TCP connections and server processes.

< STUFF DELETED >

>Has anyone tried to drive this class of machine hard in a
>"connection-oriented" client/server relationship? Is there any relavent
>performance analysis literature in this area? Or does anyone have any
>fundamental insight into this problem based on their knowledge of
>Unix?  Any comments on this topic would be welcome. If there are
>vendors that have specifically addressed these issues in their kernel
>implementation, I would like to hear about it.
>

We too would like to explore this field.  Has anyone out there
written any 'client/server' testing software?

This would be useful in determining where the limits lie for each
system and the gains and/or losses achieved from kernel tweaking.

Let's here from all of you LARGE client/server sites out there!

-brenda
-- 
% Brenda Parsons                  
% Currently at Prospect Electricity
% 10 Smith Street, Parramatta 2150, Australia
% +61 2 635 0300              e-mail:   parson@coulomb.pcc.oz.au

Re: Upper bounds of TCP & server processes in Unix

#57

Author: hschadha
Date: Sun, 08 Nov 1992 05:21

22 lines
1420 bytes

In article <Bx7JGy.6BJ@gpu.utcs.utoronto.ca>, eric@cathaus.utcs.utoronto.ca (Eric M. Carroll) writes:

|> Has anyone tried to drive this class of machine hard in a
|> "connection-oriented" client/server relationship? 

I had a tough time once last year (being then a novice in the world of
UNIX 4.3 BSD sockets) trying to figure out why my distributed matrix
multiplication program hung up for big problems (with higher number of
non-blocking calls to remote servers). Experimenting with different sizes
of matrices (and blocks), I found that the problem arose after about
fifty client-server stream (connection-oriented) connections.

The problem was in my leaving the sockets open, even after the client had
successfully demanded the result. Yes, it was the wrong way to look for 
errors. But .......

I was using TCP/IP on the DEC 2100s here.

-- 
 Harpreet S. Chadha           
 Graduate Student
 Civil Engineering, NCSU

Thread Navigation

This is a paginated view of messages in the thread with full content displayed inline.

Messages are displayed in chronological order, with the original post highlighted in green.

Use pagination controls to navigate through all messages in large threads.

Back to All Threads