for most of the fundamentals on Collaborative Filtering, You may check Chapters 8 and 9 from the following online book.

https://docs.google.com/viewer?url=http%3A%2F%2Finfolab.stan...

Shameless Microsoft. They don't have the ability to compete in the market. They just want to stop other ones' innovation.

This is because that Latency is strongly relevant to Consistency.

Clearly Ramakrishinan never very clearly clarified what Yahoo's Sherpa is. But clearly it's so much inferior to what Google has been delivering, GFS/Mapreduce/BigTable, and Spanner ...

Amazon S3 may not be readily classified as a key-value store. It seems better sit between distributed key-value store and distributed file system.

Who ignites Google's inspiration for infrastructure innovation? This man certainly should be Urs Hoelzle. All the star work, done by Jeff Dean and other at Google, is from his thoughtful planning ...

Also, does the gfs1 still have single master, so that the gfs2 has mulitple Bigtable tablets serving as distributed masters for gfs2? Is this the cause for "In fact, it just makes the bottleneck limitations of the system’s single-master design more apparent than would otherwise be the case.", as stated in the article?

gfs1 is still single master, but the workload is much simpler in this case: it serves the gfs2 master bigtable cluster exclusively. Most of the documented gfs master failures are due to misbehaved map-reduce clients. Also the gfs1 master can be down for extended period of time without affecting the master operations, due to the nature of the cluster (you're unlikely to create a million files per second resulting in much compaction and splits in metadata tablets)

The quote you mentioned actually meant that if you use Bigtable on top of gfs1, the single master failure is more apparent due to the low latency requirement of the application that use the Bigtable.

Is this vacaya related to the vacaya of hypertable? :-)

Ahaa , we had just conceived this way several months ago in China ...