According to Ambrose, the Isilon clusters have been pushed to the point of instability and additional work is being done to move infrequently accessed assets into slower storage (some of this was accomplished in 2008 - but that project is being extended). How slow is 'slower'? Perhaps as much as a second extra for the largest assets, we think. A second is an eternity in computer systems time.
The way the grid is structured there are currently three tiers of asset storage (last we looked): Assets cached locally in sims being the most rapid, and the Isilon clusters providing the next tier. Compressed assets on the Isilon clusters represent the third tier of storage, and some other unspecified system (perhaps the original asset storage clusters) look to become the fourth.
The asset cluster (or its backup cluster) isn't always at fault, however. Problems can often be traced to the utility servers, which access central storage in quite a variety of ways, but have little sense of humor when it comes to fluctuations in throughput. Differentiating problems to the point that there is a clearly identifiable fault that can be isolated to the central databases or the utility servers is no small task.
As for LLnet (freeing the grid from the vagaries of VPN servers, and reducing dependency on a single interconnect provider), there's not much detail to be had other than it is expected to be up and operating within just a few weeks.
Agent Inventory Services are expected to deploy not long after that. This is something that we've been looking forward to for a long time, and is expected to provide a great deal more speed and reliability to the overall inventory process, and nearly eliminate inventory loss and the appearance of it.
| ||Are you a part of the most widely-known collaborative virtual environment or keeping a close eye on it? Massively's Second Life coverage keeps you in the loop. |