Ian Cheong wrote:
> At 7:31 pm -0400 23/9/06, Horst Herb wrote:
>> On Saturday 23 September 2006 18:53, Ian Cheong wrote:
>>> We had this discussion on failure tolerant clusters a while back - at
>>> the time, nobody owned up to having one.
>>
>> And it might be expensive overkill.
>
> [...]
>
> What's overkill about a second server class machine and shared storage?
>
> And having fat clients that can act as servers is not overkill?
Horst probably assumed that by "failure tolerant clusters" you meant
something more specialised, as I did. But agree that what you describe
is a good idea and not necessarily overkill. Using clusters of commodity
hardware ("clusters of workstations" or COWS) is how Amazon, Yahoo and
Google achieve massive storage and processing power with extremely high
reliability. Two $2000 servers together with four $1500 workstations
configured in some redundant fashion is going to be more reliable than a
single $10,000 server from a tier-one supplier like HP or IBM.
Tim C
_______________________________________________
Gpcg_talk mailing list
[email protected]
http://ozdocit.org/cgi-bin/mailman/listinfo/gpcg_talk
