3 ms·

So each CPU has its own 256GB of local memory, with a many-TB shared pool that any of them can load/store into, using a new interconnect. How is this any diffe

So each CPU has its own 256GB of local memory, with a many-TB shared pool that any of them can load/store into, using a new interconnect. How is this any different bird's-eye architecturally from current builds involving compute nodes talking to a shared RAM-backed datastore? Is it fundamentally different to put it directly into the address space, with the collision problems that presents?

jdnier 11y ago
"to allow addressing up to 32 zettabytes (ZB)" -- I think that's the first time I've seen ZB in print. Wikipedia link: "1 ZB = 10007bytes = 1021bytes = 1000000000000000000000bytes = 1000exabytes = 1billionterabytes = 1trilliongigabytes."
- bronson 11y ago
  ZFS was originally named the Zettabyte File System. Not sure why Sun changed it. The acronym doesn't mean anything now.