Thanks Mike and John for your feedback. My understanding is that this
"mini tender" is actually still a tender in the sense that we have to
provide a clear specification of what we need, and the suppliers have to
meet these requirements. Once we have put out the tender we can no
longer discuss our requirements and iterate possible solutions with the
The advantage, as John says, is to shift some responsibility and risk to
the supplier. The downside is that we have to come up with a rigorously
specified figure of merit that we can use to select the winning bid. I
think this rules out using our own "best guess", but from Mike's
comments it sounds as if HEPSPEC figures might be available from more
suppliers than I imagined.
On 28/06/10 10:33, John Gordon wrote:
> I am sure Martin can give more detailed advice on the details you ask about but on the general issue of framework agreements we, at RAL, have been hesitant over going that route for T1 purchases. Although there are obvious benefits in reducing the overhead of tenders and in delivery, we worried about the balance of risk. In our EJ tenders we ask the supplier to meet certain requirements in benchmarking, and level of support for SL. They tender equipment that will meet our spec so they are taking some risk of us rejecting it if it doesn't. With a framework agreement we are buying from a shopping list and we make the decision on what is suitable. Our risk of getting it wrong.
> That said there is less risk with WN so we may get kit from our framework suppliers next year and benchmark ourselves. Disk servers are more difficult though so we are unlikely to go the framework route unless we can set up framework agreements with our trusted suppliers.
>> -----Original Message-----
>> From: Testbed Support for GridPP member institutes [mailto:TB-
>> [log in to unmask]] On Behalf Of Ben Waugh
>> Sent: 26 June 2010 10:06
>> To: [log in to unmask]
>> Subject: Advice on procuring worker nodes
>> Dear All,
>> This is an appeal for guidance from both those who have put out
>> tenders for CPU nodes, and those with knowledge of what makes a good
>> worker node for ATLAS in particular.
>> I know procedures vary between institutions, but I have been advised
>> by our Procurement department to do a "mini tender" involving the five
>> suppliers who have framework agreements to supply servers to UCL,
>> asking for the greatest possible CPU power for a fixed price.
>> The HEPSPEC rating is the obvious measure to maximise, but not all
>> suppliers have the means or inclination to run a specialised benchmark
>> for a relatively small order, about £40k. How have others done this?
>> Do you restrict yourselves to the suppliers who already have
>> experience in dealing with GridPP and can run HEPSPEC themselves, or
>> do you use other benchmarks or some less direct way of comparing the
>> CPU rating of the products on offer?
>> There are of course other factors affecting job throughput, including
>> hard disks and RAM. Is there some way of measuring the effect of
>> these, or would you just set a minimum requirement on both and then
>> maximise the HEPSPEC? If you would take the latter approach, what is a
>> sensible trade-off between disk performance and price? Presumably
>> 10kRPM SAS disks will be better than 7.5kRPM SATA, but maybe a striped
>> pair of slow disks would be an alternative? And how much disk space do
>> you allow per CPU core?
>> If there is anything else I haven't asked but you think I should
>> consider, please tell me that too!
>> Best regards,
>> Dr Ben Waugh Tel. +44 (0)20 7679 7223
>> Dept of Physics and Astronomy Internal: 37223
>> University College London
>> London WC1E 6BT
Dr Ben Waugh Tel. +44 (0)20 7679 7223
Dept of Physics and Astronomy Internal: 37223
University College London
London WC1E 6BT