[Serdev] SER's core design features(process model/parser/lumps/script) - was: So who/what is SER for, anyway?

[Martin Hoffmann]

Dragos Vingarzan wrote:
>   
> But Martin, telco's toys come with that thing built in. And that is a
> different approach from the Internet-like one.

Indeed. What you have here is a clash of philosophies. In Telco land you
happily pay insane amounts of money for a complete solution. In Internet
land you get a toolbox and, if you are lucky, a manual to build your very
own solution. Personally, I prefer the latter. But then again, that's
where I come from.

> Their toys are built to
> survive and work even if the sysadmin decides to take a long vacation
> and several things crash. They have things like 3 redundant nodes, with
> total load of max 66%, so that in case that one fails, the others can
> take over. And all this is BUILT IN and TRANSPARENT, like nothing
> happened. No BOFH needed ;-).

I have more or less the same. Okay, it didn't come in a box, I had to
build it myself.

> If the IMS convergence thing will ever work, then the telcos will learn
> some lessons on "freedom" and "openness". It would just be a shame if
> the IP guys won't learn the reliability thing back from them. I worked
> for some time in parallel with guys doing that and believe me that they
> were more serious than we are now.

Seems to be the other way around, judging from what the telcos guarantee
for DSL. Thing is, reliability has a price tag attached. You can do
plenty of nines in a large network if you have the quasi-monopoly. In a
competitive market, you have to make decisions.

Regards,
Martin