AMD scores EPYC gig powering new Azure instances

Don't pop the champagne yet: Microsoft's still using Intel in same series

4 Reg comments Got Tips?

AMD's return to the server market has won it a place in a top tier cloud: the company today revealed that it has won Microsoft's business for the next generation of Azure's L-series instances.

The chip-maker told The Register its 32-core, 2.2GHz EPYC 7551 will power a new “Lv2” Azure instance type. Azure's L-series instances are optimised for storage and workloads like databases and Apache Spark.

AMD's representatives positioned the win as a triumph that shows the firm is a force in the server CPU market again.

But the company shouldn't pop the champagne just yet.

To understand why, know that the EPYCs will be used in the Olympus, an Open Compute server design that Microsoft has started to run in its Fv2 instances.

Microsoft has previously said that Olympus servers would also run Intel's FPGA-equipped Xeons. The Fv2 instances also run a “cryptographic microcontroller” co-created with Intel that secures the firmware of devices on the servers' motherboards.

Throw in the facts that AMD would not comment when asked how many EPYCs Microsoft has acquired and that the new instance type is in preview only, and it looks like AMD has a foot in the door rather than an invitation to dinner.

However, the company has foreshadowed it would soon reveal other wins with substantial clouds. At a press event in Sydney, Australia, on Monday, the company's director for world-wide commercial client business development, John Hampton, said AMD aspires to once again enjoy 25 per cent server market share. ®


Biting the hand that feeds IT © 1998–2020