Storage
Alacritech apprehends an NFS anomaly
Alacritech claims NFS filer I/O is grossly skewed towards reads and suffers from read metadata processing that chokes controller CPUs. It has just launched its ANX 1500 filer accelerating cache product based on its recognition of NFS read metadata filer I/O loads that can overwhelm filer processors and delay file delivery. A …
The v3 metadata problem is common knowledge
V3 metadata reads can spoil even a small network (not just enterprise).
Classic example - firefox/thunderbird. These two pieces of software becomes a network DOS tool if the users $HOME is on NFS v3. They use SQL-lite (which is increasingly popular with many other developers as well) to maintain most of the cache, history, etc. The result is that even one user on a 1Gbit network can produce a rate of metadata requests that is enough to choke a rather reasonable NFS server to a point where you cannot read a basic DVD SD movie off it without interruptions.
"Accelerating" such metadata does not fix the fundamental problem with v3. The more apps use embedded databases like SQL-lite the worse it will get and as long as it is v3 there will be more and more requests for metadata to filers.
The real solution to this is NFSv4 which allows caching some of this metadata locally and telling the server that it has been cached (something SMB/CIFS has been doing for 10+ years now).
just curious...
But is this a sales pitch/advertisement or an IT story?
It reads like a press release more than anything else...
Analogy issues.
"...a stupid idea, like trying to combine car engine super-charging and turbo-charger...."
Such a stupid idea that it's in fairly common use. A small supercharger provides boost at low revs where the turbo doesn't operate and a big turbo provides efficient boost at higher revs. It's one of the ways of getting round the dreaded "turbo-lag" and has the advantage that both units can be optimised for the boost range they're dedicated to for maximum efficiency.
A bit like combining filer metadata and raw data caching systems rather than trying to shoehorn the two different functions into one product. Maybe it's not such a bad analogy after all........
It's marked "comment",
but all of the commenting seems to've been done by various marketeering veeps. The main question I have, and which the article leaves unanswered, is "why am I reading this?" To watch a Battle of the Press Releases? Would that fly on television?
