Re: Persistence & Cache clusters

To: Lars Marowsky-Bree <lmb@xxxxxxxxx>
Subject: Re: Persistence & Cache clusters
Cc: lvs-users@xxxxxxxxxxxxxxxxxxxxxx
From: Wensong Zhang <wensong@xxxxxxxxxxxx>
Date: Tue, 28 Sep 1999 23:38:04 +0800

Lars Marowsky-Bree wrote:
> I think the real servers SHOULD NOT require a special modified kernel, since
> they may not be running Linux.
> The kernel on the LVS needs to replay the connection establishment to the real
> server who should never notice that anything went wrong. After the replay, it
> is "just" basic NAT, fiddling with the packets on the way out again (changing
> the source address, adjusting the TCP sequence number etc).

Yeah, it probably works with NAT, but it cannot be applied to 
the LVS/TUN or LVS/DR, because the real servers send response
packets to the clients directly.

> Of course, application level scheduling is a way more powerful tool than just
> that.

Application-level scheduling is easy to implement, but its
overhead is high, because there are two TCP connections for
every request, passing packets to the application and passing
back increase the delay.


---------------------------------------------------------------------- mailing list - lvs-users@xxxxxxxxxxxxxxxxxxxxxx
To unsubscribe, e-mail: lvs-users-unsubscribe@xxxxxxxxxxxxxxxxxxxxxx
For additional commands, e-mail: lvs-users-help@xxxxxxxxxxxxxxxxxxxxxx

<Prev in Thread] Current Thread [Next in Thread>