IBM has refused to comment on a damning document from consultancy Xephon, attacking the performance of the supplier's Parallel Sysplex mainframe clustering.
This is the second time Xephon has seriously questioned Parallel Sysplex. On 14 April, IBM published a document outlining its strategy after Xephon criticised it for ?overheads?, showing that for every 100 Mips of processor capacity, a user in the data sharing environment will need to buy between 141 and 209 total Mips and pay a software bill based on 125 to 167 Mips.
The author of the report, consultant Barry Graham, said in his report: ?The Parallel Sysplex can never be cost competitive. Even when capacity needs exceed the single largest system the economics should be studied carefully.?
Graham described IBM's performance data as "dubious". ?Even at the 20-processor level, the Parallel Sysplex is inferior,? he continued.
Xephon has other major issues with IBM?s system. He said: ?The coupling facility Mips used in the tests were on average well over 35 per cent of the Mips of the systems being measured compared to IBM?s recommended level of 10 per cent. The tests were run at utilisation levels below 100 per cent and the expected performance at 100 per cent loaded was extrapolated. Response times more than doubled in some cases.?
Equinox's Dave Millett explores how phone, mobile and broadband could be affected by a no-deal Brexit
Dust storm on Titan only the third Solar System body where such storms have been observed
New technique could enable quantum computers to scale-up to millions of qubits
Systrom and Krieger taking time off "to explore our curiosity and creativity"