activemq-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jim Lloyd <jll...@silvertailsystems.com>
Subject Re: Compression on network connections
Date Tue, 06 Apr 2010 15:55:40 GMT
On Mon, Apr 5, 2010 at 6:57 PM, Bruce Snyder <bruce.snyder@gmail.com> wrote:

> On Mon, Apr 5, 2010 at 2:48 PM, Jim Lloyd <jlloyd@silvertailsystems.com>
> wrote:
> > I'm in the process of building a network of brokers in which some of the
> > links are geographically distributed. These links will use the SSL
> transport
> > for reasons of security, but we also want to use compression to minimize
> > bandwidth usage over the WAN links. I've searched the available
> > documentation and so far haven't found a definitive source that tells
> > whether the SSL Transport uses rfc3749 compression or not. Can anyone
> here
> > give me a definitive answer? If it is supported, which configuration
> options
> > are required?
>
> Not only is TLS compression is not supported in ActiveMQ, but I don't
> think it's supported in Java yet. Do you know of an implementation?
>

No, I don't, but we do the majority of our development in C++ so even though
I hadn't found an implementation in my searches so far I still had hope that
one might exist. Thanks for the clarification.

>
> > It also appears that there is an option to use message level compression.
> If
> > compression is available on the SSL transport stream, I expect we would
> want
> > to ensure that message compression is turned off to allow the transport
> > level compression to have more context. The only documentation I have
> found
> > for this was for ActiveMQ 3, which implies that doMessageCompression
> > is trueby default. Is that still the default with ActiveMQ 5?
>
> Message level compression can be enabled on a ConnectionFactory or a
> Connection via the useCompression property. See also the
> useCompression option on the URI:
>
> http://activemq.apache.org/connection-configuration-uri.html


Thanks. I see the default for useCompression given here is now false. We'll
experiment and see if this gives us any significant improvement.

>
-Jim

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message