Object Databases to NoSQL and to Cloud Databases

Software industry is characterized by changes-rapid changes. New technologies get introduced on regular basis. It is also characterized by technologies getting morphed and present themselves in new avatar or get re-purposed.

Take example of object databases. Back in 1980, objects database management systems(ODBMS) were considered serious contenders to relational alternatives. They out-performed RDBMSes in many situations. It allowed application developers think in terms of objects end-to-end, without any impedance mismatch. They alleviate costly joins processing and provide consistent persistence. Many companies existed and still exist even today. I used to work one such company called Versant. In fact, before that, I worked for another company called MediaDB which was based on ODBMS, but was built and optimized for storing multi-media objects. Versant, was and is general purpose ODBMS like Objectivity, O2, Poet, GemStone, etc.

ODBMSes built their businesses by providing data management solutions for complex application having complex object graphs and need to treat them end-to-end that way. ODBMSes became prevalent in niche market and were never able to make to big league for variety of reasons. They often struggled to justify their existence. Then came two waves of technologies which made them relevant once again. One of them is NoSQL wave. And the other is Cloud wave.

With proliferation of Internet and web based applications, need of managing unstructured data became need of the hour. Big data analytics added further spice to how data from variety of sources is treated, managed. SQL was no more perfect for such applications, hence many open source initiatives brought NoSQL movement and came up with data stores with based on non-relational concepts. ODBMS vendors quietly watched this revolution. ODBMSes were the first NoSQL databases. As NoSQL databases became more prevalent, ODBMS vendors meta-morphed themselves and requisitioned to ride on NoSQL movement. Look at Versant’s website to see the case in point. They, of course, got acquired by Actian recently. But this re-positioning of Versant began much before acquisition, via technologies such as Versant JPA.

With proliferation of cloud and wider adoption of cloud, along with new age applications such as big data analytics, ODBMSes are well positioned to apply themselves to these needs and to the scale of cloud. Many cloud vendors do provide object storage capabilities such as Amazon/AWS S3 or even open source Object Storage initiatives such as Ceph or Swift which are can be used with OpenStack or CloudStack. But they are not ODBMSes. They are raw object storage mechanisms and very useful. To cater the needs of second wave, ODBMSes have started to position themselves for cloud. Cloud computing means scale, high volumes, distributed databases, and typical applications are analytical in nature than more of transactional. For ODBMSes, there are few facets of cloud database which needs to be still taken care of such as Database as a Service(DBaaS), high availability(HA) from cloud ecosystem perspective. We will see these gaps from cloud support perspective, being addressed in coming years.


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