Radoslaw Skorupka writes: >* RAIM as memory redundancy is not unique to mainframes.
It is, actually. More precisely, nobody else seems to have done anything except "Level 1" mirroring(*), rarely even that, and even then not always/usually in hardware. It's a question of how many hard failures are tolerable. On zEnterprise servers (single machine, z196/z114 and newer) as many as 3 are tolerable in the memory subsystem. The number 3 is greater than the number 1. It's entirely fair and accurate to say this stuff is qualitatively different because in fact it is. That's only one way it's different. Yes, you have a choice if you don't need or want the world's best server memory subsystem: buy another server (or a hundred). IBM, for example, offers a variety of servers with a variety of attributes. (*) zEnterprise servers support opportunistic memory mirroring, in hardware. That's an option *above* RAIM, and it's available primarily if you want to plan for (unlikely) processor book maintenance events. Look for "Flexible Memory Option" and "Enhanced Book Availability" in the zEnterprise specifications and documentation. zEnterprise servers also support selective memory mirroring (via the Coupling Facility). -------------------------------------------------------------------------------------------------------- Timothy Sipples IT Architect Executive, zEnterprise Industry Solutions, AP/GCG/MEA E-Mail: [email protected] ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
