Re: 10,000+ Simultaneous Users

From: Alan <alanshein_at_spambuster.erols.com>
Date: Thu, 22 Mar 2001 09:58:22 -0500
Message-ID: <99d3sn$5ha$1_at_bob.news.rcn.net>


Thanks. Answers in-line.

"Jason Burgoyne" <jason.burgoyne_at_bms.com> wrote in message news:3AB8FD94.25C71770_at_bms.com...
> I really only know unix, but I know to give you a better answer we will
> need the following info.
>
> LAN or WAN?

Internet

> Thin client or Browser?

Browser

> Expected response time?

20 seconds or less

> Expected response data, web page, XML, raw data, formatted report?

formatted report

> How many and how big are the denormallized tables, are there some that can
> be pinned into mem?

Don't know yet. I suspect we will eventually wind up reporting agianst a star schema, so pinning tables is not an option, as they will be too large.

> Is there a complex authentication system or can anyone get everywhere?

Only a couple of types of users who will have access to different schemas.

> If you are using a WAN is replication an option to distribute the load to
> individual LANs?

n/a

> If so would you need MASTER-MASTER replication?
>
> There are a lot more, but I think those are the ones that add signifactly
> to the cost.
> JB
>
> Alan wrote:
>
> > What do I need in the way of hardware to support 10,000+ simulataneous
> > users, given the following:
> >
> > 1. OS must be NT (Windows 2000) as opposed to Unix.
> > 2. Each user's processing is not expected to be intense. They will be
> > running canned queries against denormalized reporting tables.
> >
> > Thanks.
>
Received on Thu Mar 22 2001 - 15:58:22 CET

Original text of this message