Saturday, December 12, 2009

Will NoSQL Rescue the World Of Data in the Cloud? (Info and Links for the Tech Curious)

In a week, I will have three weeks off from work, and since I usually give in to my inner geek when I'm on vacation,  I thought I'd find out more about cloud computing and NoSQL today.   I was overwhelmed by what I found within the first minutes of my search!  (It has been about 4 years since I took a database class.)

The first place I visited was the O'Reilly Radar blog, which usually has a few interesting links about emerging technologies:  Four Short Links:  11 December 2009  (Nat Torkington)

Number 2 on  Torkington's list of links:
NoSQL Required Reading   "Papers and presentations to get up to speed in the theory and practice of scalable key-value data stores" (via Hacker News)

Here is a sample of some of the info I found during my first trip down the NoSQL rabbit hole:

Braindump Blog
NoSQL Debrief,  (Johanos Karsson)
"The idea was to give attendees a solid introduction to how distributed, non relational databases work as well as an overview of the various projects out there." (The NoSQL DeBrief includes a variety of presentation slides and videos that I'm planning on viewing on my next tour of the rabbit hole.)

Pragmatic Programming Techniques  (Ricky Ho's blog)
Query PRocessing for NoSQL DB
       "The recent rise of NoSQL provides an alternative model in building extremely large scale storage system. Nevetheless, compare to the more mature RDBMS, NoSQL has some fundamental limitations that we need to be aware of. It calls for a more relaxed data consistency model.It provides primitive querying and searching capability."
       "There are techniques we can employ to mitigate each of these issue. Regarding the data consistency concern, I have discussed a number of design patterns in my previous blog* to implement system with different strength of consistency guarantee."
*NoSQL Patterns

All Things Distributed - Werner Vogel's weblog on building scalable and robust distributed systems: Eventually Consistent- Revisited



Cloud Computing:  Ground Computing, Local to the UserCouchDB - Local Web Platform

View more documents from Chris Anderson. (Interesting visual on the slides. Chris has a good sense of humor.)

Video Version from YouTube Google Tech Talks:  Chris Anderson Discusses CouchDB

Chris is "Obsessed with bending the physics of the web to give control back to the user".

Dave Rosenberg's Software, Interrupted Blog,  CNET News 12/9/09

Update 12/12/09

Here are a few more links, thanks to Nati Shalom:

Why Existing Databases (RAC) are So Breakable!  Nati Shalom (11/30/09)

No to SQL?  Anti-database movement gains steam- My Take Nati Shalom (7/9/09)

Are Cloud Based Memory Architectures the Next Big Thing?  Todd Hoff (3/16/09)

Other alternatives:  In-Memory Data Grid


I hope this won't be another "boys club" thing.  The NoSQL community doesn't appear to have many representatives from an important gender that will increasingly rely on cloud computing technologies during their daily lives.

I stand corrected.
Delegates at the Berlin NoSQL meeting- there looks like there are at least three women in the group:

NoSQL pic1
Monika Moser was a presenter at the NoSQL conference and was featured in a blog post written by Isabel Drost and Jan Lehnardt:  Happenings:  NoSQL Conference, Berlin

"More and more web applications have data storage requirements that can't be fully met using traditional relational databases. Object-oriented and document-oriented databases provide an alternative – and a field of development that has recently seen much activity."

According to the information on post, Monika developed the transactional storage layer of Scalaris, and is currently an Erlang/Ruby/Hadoop developer at

NoSQL pic2


Nati Shalom said...

There are various reports that was gathered from data center statistics that analyze disk failure statistics:
* Actual disk failure/year is 3% (vs. estimates of 0.5 - 0.9%) – this is a 600% difference on reported vs. actual disk failure.
* There is NO correlation between failure rate and disk type – whether it is SCSI, SATA, or fiber channel. * * There is NO correlation between high disk temperature and failure rates

Those analysis shows that the approach of relying on a shared storage for reliability as with most of the existing database RAC clusters is broken. Instead NOSQL alternative approach assumes that failure are inevitable and where designed to deal with those failure even under extreme scenarios.

I summarized that topic on one of my recent post: Why Existing Databases (RAC) are So Breakable! and here

You may also want to add another category of NOSQL alternatives: In-Memory-Data-Grid

See more details on that regard on Todd Hoff ( write-up: Are Cloud Based Memory Architectures the Next Big Thing?

Lynn V. Marentette said...

Thanks for the information and links, Nati. I updated this blog post to reflect your contribution.