User services (was: Re: dict server)

On Sun, 2006-01-22 at 16:25 -0200, Guilherme de S. Pastore wrote:
> Em Dom, 2006-01-22 �18:04 +0100, Reinout van Schouwen escreveu:
> > Hello all,
> Hi,
> > 
> > Following up on bug 167366, I am hereby asking if it would be possible to run a 
> > dictionary server on
> [snip]
> I personally think that our current infrastructure cannot handle any
> more services, specially as widely/frequently used ones as a dictionary
> or weather service (and I am not sure it will ever be able to host such
> a thing). And for their being publicly available, I am not sure if
> would be the correct place to host them (at least as long as
> we have pending improvements for our own services).
> GNOME has a huge number of users, and even non-GNOME users might start
> using our service if it is good too, which will almost certainly
> overload us. Is there a way we could host only a list of dictionary
> servers? That way I think it might be closer to something feasible.

I've had this message tagged for a while, because I thought it was worth
a reply.

I think Guilherme is right that we can't add more services to our
current servers. The problem is not so much one of capacity, but rather
that all 4 of our servers are already running many different services,
and piling more onto them just increases the probability that something
will take down the box and we'll have trouble figuring out what.

We saw considerable increases in stability going from 1 server to 2 
servers (splitting CVS from web), we saw further increases in stability
when we split off mail, and yet further increases when we went to
the current 4 server setup (adding a database server).

But the set of servers we have isn't fixed; we have 3-4 more servers in
the pipeline to be in place within the next couple of months.

And I disagree with the larger idea that we shouldn't be running
end-user services. What percent of time that you spend on the computer
is spent not using network resources? No web, email, cvs, online games,
etc. Probably close to zero. If GNOME defines itself out of the
networked space, than it has basically defined itself out of the space
of current computing and innovation.

While there is the ability to piggy-back on existing services,
and some ability to bypass the need for services altogether by 
going peer-to-peer, we should not make things hard for ourselves
and unpleasant for the user by trying to avoid needing a centralized

I don't think we need to be scared of the idea that we might swamp
our limited server capacity (even 8 servers is a pretty small set),
or swamp Red Hat's bandwidth. If we create things that are genuinely
useful and interesting to people, it's easy to fund-raise for new
servers, and it's easy to find people to host bandwidth. Swamping is


(within reason :-)

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]