Life with ‘NoSQL’…

I  have to say that after playing with Cassandra, a NoSQL datastore, I’m completely impressed with what it gives. Fast read and writes, scale unlike I’ve ever seen, and (fairly) easy to manage. Its not without compromises; transaction management, two-phased commits, atomicity, etc. But I’m having to view the datastores in a whole new light. Here is a list of interesting links of talks and presentations:

The last one contains a bunch of links to videos with various NoSQL solutions. I highly recommend to go though them.

At this point, I can see these NoSQL datastores to be used along-side SQL ones. There is no question that many websites can only use NoSQL datastores, but it gets tricky when dealing with things that SQL databases solve well, such as financial transaction management. Considering that many large banks still use datastores like IDMS, that could eventually change.

Advertisements

One thought on “Life with ‘NoSQL’…

  1. Gopi Nathan says:

    It is nice to see discussion on NOSQL databases. Nothing new here except the new products. Concept was there and even many products were there – all on IBM mainframe.

    In the 1970s and 80s discussions on databases were centered around hierarchical model, network (CODASYL) model and relational model. IMS,IDMS and System-R prototype examples of each. Later in the mod 80s IBM gave an implementation caleld SQL/DS on VM/CMS, leading finally to DB2 on MVS. Oracle, Ingres and later Sybase/Informix followed on VAX and Unix platforms. MS/SQL is an offshoot of Sybase.

    Relational databases though alright for many applications, failed to provide OLTP support for large data. And in recent times, Javaworld started complaining about the lack of support for Objects / Domain concepts. Java Objects were cooked up from relational databases. This led to products supporting ORM – Object-Relational mapping… a sort of re-inventing the wheel – creating CODASYL style E-R structures from pure relational tables…

    Products like IDMS are no longer based on pure CODASYL sets though literature is not updated on this. IDMS provided full support for idnexes as early as 1984 and I know customers who made databases without involving a single ‘SET’. IDMS also supported SQL access against NONSQL (sic) database. It also supported pure SQL tables.
    But none of these seem to be noticed by the industry analysts and writers!!!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s