<< | Thread Index | >> ]    [ << | Date Index | >> ]

To: cipe-l,AT,inka,DOT,de
Subject: Re: Speed Question
From: Richard Gregory <R.Gregory,AT,liverpool,DOT,ac,DOT,uk>
Date: Fri, 13 Feb 2004 20:11:53 +0000
In-reply-to: <402BB3FF.1070401@oec-sys.com>
References: <402BB3FF.1070401@oec-sys.com>

Hi Bruce,

There are a few things to check to locate this problem. First, what
machines are running the CIPE end points? Encryption takes computational
effort, the endpoints must have enough power to cope with the maximum
bandwidth.

Also, a problem I encounter is the CIPE packet overhead. CIPE adds 54
bytes or so to each packet, so if the packets are small and there are
many of them the overhead becomes significant. To test for this you can
compare the datarate going through the CIPE interface to the datarate
through whatever the CIPE traffic runs over.

Somebody else on the mailing list was asking similar questions. He was
wondering why the datarate over CIPE varied so much in the up and down
directions and between protocols. Protocols like FTP had more throughput
than NFS. The high speced CIPE end point had higher outgoing throughput
than the lower speced endpoint. The differences were significant.

Richard

Bruce Sackett wrote:
I am running into what appears to be some slow connections across multiple CIPE connections. Ping rates are extremely good, but data transfer across these connections is reportedly very slow. Most transfers across these connections are SQL queries and responses at this point, with some file transfers. In many cases these queries are quite large, or are many many small ones.

Any possible suggestions as to what I can check on this, or settings/options I should look at?




<< | Thread Index | >> ]    [ << | Date Index | >> ]