A reason #meltdown is so important now, and possibly why it wasn’t discovered earlier in the public, is only recently have our environments shifted significantly to co-tenancy (AWS, Azure, GCP...). The attack is meaningful due to target change.
Think big picture.
1/2
-
Show this thread
-
Replying to @dotMudge
I'd say "shifted *back* to co-tenancy" -- back in the mists of time, many companies used timesharing service bureaus.
1 reply 0 retweets 15 likes -
Replying to @SteveBellovin @dotMudge
Have we officially (sp correct this time) looped back around to MULTICS?
1 reply 0 retweets 4 likes -
Unfortunately not. Systems running on MULTICS didn’t perform lookahead speculative execution ;-) Time sharing systems still exist, but they’re multi threaded and single ‘program’ use rather than multi user.
1 reply 0 retweets 2 likes -
Replying to @pentestmatt @_CastleinthesKy and
That's this bug, but might there have been others? The 360/85, circa 1968, had cache, instruction and operand prefetch, and overlapped execution (http://www.bitsavers.org/pdf/ibm/360/funcChar/A22-6916-1_360-85_funcChar_Jun68.pdf …).
3 replies 0 retweets 6 likes -
Replying to @SteveBellovin @_CastleinthesKy and
Oh, gee. ha. But no speculative branching. :-D
1 reply 0 retweets 0 likes -
Replying to @zeynep @_CastleinthesKy and
I truthfully do not recall. I have vague recollections of the 360/91--a supercomputer for its day--having it, but it didn't have cache except for a very small instruction cache for "loop mode".
@Koenig_Software, do you remember?1 reply 0 retweets 0 likes -
Replying to @SteveBellovin @zeynep and
360/91 had out of order execution, which implies prefetch, but I I’m not sure of its branch prediction capability.
1 reply 0 retweets 0 likes
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.