RE: [logs] RE: syslog/tcp (selp)

From: Rainer Gerhards (rgerhardsat_private)
Date: Wed Jan 15 2003 - 01:38:01 PST

  • Next message: Bennett Todd: "Re: [logs] RE: syslog/tcp (selp)"

    > > I agree that "event log" sounds a little "Windowish" and maybe 
    > > misleading. So I, too, would like to switch to SLP name.
    > I'd look for something pronouncable, but defined, SLP could 
    > be SLIP or 
    Oops... Good warning sign. I tried to find something that is not in use
    yet (at least widespread). I have seen so much terms used for totally
    different purposes, that I try to not add any confusion. And SLIP is
    "Serial Line IP", so it could cause some...
    > SLOP- I also think "simple" was overloaded with SNMP, which 
    > isn't all that 
    > simple.  I know it's sort of marketingish, so I've not copied 
    > the list 
    > (feel free to throw stuff back on-list if appropriate).
    Well, I think this kind of "marketing" is important. After all, we would
    like to see it used. So I am posting this reply to the list.
    > I'd go with 
    > Reliable <something>, so that it's selling itself to 
    > developers as a good 
    > thing.  In the non-pronouncable space, RLP would probably be 
    > better than 
    > SLP.
    Honestly, I do not like to call it reliable as it is not fully reliable.
    RFC3195 is. But SELP/SLP/RLP ;) is a simplex protocol, without any
    acknowledgment from the receiver (except for the TCP acknowledgment).
    This has some inherent issues in it. Sure, it is much more reliable than
    syslog/udp, but not bullet proof (maybe 90% to 95%, which is not bad at
    all). So it could be MRLP (mostly reliable log protocol) - but hold on,
    I am just kidding... ;)
    I'll address the reliability issues I see in the upcoming next version
    of the spec.
    Having said this and thinking about your comment on pronounciation,
    maybe its not bad to stay with SELP... (which still has the simple issue
    ;)). As it looks, choosing the name is the hardest part as always in
    LogAnalysis mailing list

    This archive was generated by hypermail 2b30 : Wed Jan 15 2003 - 06:24:41 PST