Roland Pope wrote:
>
> This means that the CYRUS server box doesn't
> incure the overhead of having to negotiate 2 TCP conenctions for every 1 client,
> as well as the SSL negotiation overhead.
Do you think I could resolve this satisfactorily by simply
adding a couple more processors to the box?
Roland Pope wrote:
>
>
>>would split your users to multiple servers to start with. Or use
>>front ends which is the TLS endpoint, and proxies the session to the
>>right server, then the mail server doesn't need to do TLS as well.
>>
>>
> I agree. We use stunnel on seperate machines to provid
> Gary Flynn <[EMAIL PROTECTED]> writes:
>
> > I'd be interested in hearing about others' experiences on
> > the impact of stunnel or SASL on server resources. Any
> > thoughts on the relative merits of either architecture
> > of providing SSL sessions would also be appreciated. We'll
> > need t
Gary Flynn <[EMAIL PROTECTED]> writes:
> I'd be interested in hearing about others' experiences on
> the impact of stunnel or SASL on server resources. Any
> thoughts on the relative merits of either architecture
> of providing SSL sessions would also be appreciated. We'll
> need to protect both
I'd be interested in hearing about others' experiences on
the impact of stunnel or SASL on server resources. Any
thoughts on the relative merits of either architecture
of providing SSL sessions would also be appreciated. We'll
need to protect both the IMAP and IMSP sessions.
Some background:
1)
