working with integrators
information and established relationships with the manufacturers,’ adds Stolyar who goes on to comment that knowledge of the technologies available from vendors and how they fit together is a must, as, while one customer may not benefit from having one a specific product implemented, another may see a massive performance boost. He therefore advises that end-users find an integrator that is able to fully develop an architecture, rather than simply provide hardware. ‘Quality is also a big factor,’ he continues, ‘and some companies work from a small office and don’t have the proper processes in place to deliver a quality system.’
Ensure your integrator has the expertise and resource to back up their system offering with a range of vital services. We’ve encountered numerous integrators who claim to offer support packages, warranty packages, maintenance packages, etc but fall short when delivering these services.’
Measuring the performance After the choice of integrator has been made, there is one important enquiry a customer should make to ensure that after they take delivery of the solution they have something to measure the success or failure by; they need to ask for a sufficient level of benchmarking,
IT CAN BE UNREASONABLE FOR A CUSTOMER TO ASK
FOR BENCHMARKS TO BE RUN ON EXACTLY THE SAME SYSTEM THAT THEY’RE PURCHASING BECAUSE VERY OFTEN THEY ARE BUYING THE ABSOLUTE LATEST TECHNOLOGY
Manoj Nayee echoes that thought: ‘Be
sure you understand what the integrator is offering and ensure it actually meets the requirements in the most efficient manner. We see many integrators (all the way up to Tier-1 in fact) offering less than optimal products from their standard ranges simply because they do not have the means to provide bespoke, optimised solutions.
offers Fielden. ‘The simplest benchmarks are ones like Linpack, but they aren’t “real world”. Ideally users should have a piece of their own code that they want to be run on equipment that is representative of the system they are buying. It can be unreasonable for a customer to ask for benchmarks to be run on exactly the same system that they’re purchasing because very often they are buying
the absolute latest technology. Because of this there may not be a huge benchmark system anywhere in that configuration.’ Although not a trend as yet, companies
like OCF are seeing a level of interest in ‘HPC on-demand’ with customers looking to see whether or not this makes more economic sense than the investment necessary for an on-site system, says Fielden. In November 2010, and in a first for the UK, OCF announced its own HPC on-demand service, enCORE. As a result of the company’s strong relationship with Daresbury Labs, a UK government-funded laboratory that provides support for chemistry codes and multidisciplinary research, the two entered into a partnership whereby Daresbury labs purchased equipment and OCF have a resale agreement to market the facility. Fielden concludes with the benefits of
entering into such an arrangement: ‘This provides Daresbury with a secondary source of income as, instead of buying their own systems, local SMEs or Universities can purchase time on the one belonging to Daresbury. It is currently in a pilot phase, but if it takes off as expected, it will provide Daresbury with a reinvestment model. UK public funding is somewhat under pressure and this could be a way of generating more revenue that they can then reinvest in their own infrastructure.’
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48