|
20 December 2001
Derek Prior
Alan Mac Neela
Donna Scott
Josh Krischer
Ian Brown
Jonathan Green-Armytage
Management Summary
Enterprises face many challenges in identifying and understanding all the components of hardware, software and services that are required in a large, mission-critical SAP application. All the components need to be integrated and this leads to a complex infrastructure, even for a small project.
Gartner has assembled the required elements into a model that covers, in detail, the technology and services that such an application needs -- server and storage hardware, system software, and consulting and support. Enterprises can use the model to evaluate vendor offerings for SAP projects where high availability is a priority.
As an example, Gartner specified a mission-critical high-availability system and invited Hewlett-Packard (HP), IBM and Sun Microsystems to submit proposals for a suitable system. We then scored their proposals and responses to evaluate them against the criteria in the model.
Gartner conducted this exercise at the end of 2000. The scores would have been different a few months later. By late 2001, all three vendors had introduced new high-end servers. And Sun had changed to Hitachi as its supplier for storage, which would have changed our ratings of its storage capabilities. But the methodology is more important than the result. And the dynamics of the industry mean that leading vendors will frequently change their relative positions in any ranking or selection process.
In this instance, we started by grading the vendors' responses with up to 10 points for each of 32 criteria. The methodology is designed to allow enterprises to weight the scores according to their own requirements so that, for instance, they could give more weight to proactive support services than to operating-system partitioning. Applying representative weightings for the relative importance of the criteria, the maximum possible score was 670, and we scored the vendors as follows: HP 409, IBM 355 and Sun 283.
Scoring performance against detailed...
|