Analysis

Is IMS a NoSQL database?

Philip Howard By: Philip Howard, Research Director - Data Management, Bloor Research
Published: 20th December 2012
Copyright Bloor Research © 2012
 Logo for Bloor Research

For my Christmas Jollity this year I thought I would address the question of whether IMS is a NoSQL database. And, for that matter, whether the same applies to IDMS and Adabas.

In order to answer this question we have to ask what the definition of a NoSQL database is? According to Wikipedia "in computing, NoSQL (commonly interpreted as "not only SQL") is a broad class of database management systems identified by non-adherence to the widely used relational database management system model." Well, if that's the definition then IMS definitely is a NoSQL database.

It goes on to say that "NoSQL databases are not built primarily on tables, and generally do not use SQL for data manipulation." Well, IMS is not built on tables and you don't generally use SQL to access it.

Further, "NoSQL database systems are often highly optimized for retrieval and appending operations and often offer little functionality beyond record storage (e.g. key-value stores). The reduced run-time flexibility compared to full SQL systems is compensated by marked gains in scalability and performance for certain data models". Clearly, IMS is not optimised as described - it is optimised for transaction processing - but note the "often", which implies not always and the "certain data models" which implies the same thing. Not all of what we might normally think of as NoSQL databases follow this description and thus this cannot be part of any definition.

Bottom line: according to Wikipedia IMS is a NoSQL database.

Searching for further clarity I have turned to www.NoSQL-database.org. It defines NoSQL as follows: "Next Generation Databases mostly addressing some of the points: being non-relational, distributed, open-source and horizontally scalable." That's hardly a definition. "Some" of the points indeed! I would particularly disagree with the open source categorisation: Acunu and MarkLogic, for example, both offer what are clearly NoSQL databases and neither of them is open source. Indeed, the plethora of open source companies in the NoSQL space is starting to remind me of the dot.com bubble - there are far too many companies in this area with far too much hype - some (a lot?) of the VC companies backing these open source companies are going to get burned.

Returning to this definition I don't know how "next generation database" is defined but clearly IMS isn't. On the other hand the site classifies various databases by type (much more accurately - though still not perfectly in my opinion - than Wikipedia by the way) and right at the bottom of its listing it has an "unresolved and uncategorized" list. This list includes Adabas and as a final word, it states that "VSAM by IBM is also a good candidate". If VSAM qualifies then so, surely, must IMS.

Frankly, this is daft. If VSAM, Adabas and IMS qualify as NoSQL databases then the term NoSQL has no useful meaning. As I have mentioned in previous articles, graph databases and triple stores are NoSQL databases. And since DB2 has a triple store, that technically makes it a NoSQL database!

Reader Comments

Posted: 20th December 2012 | By Peter Abrahams :

Philip
As I installed one of the first IMS systems in the UK I just had to read your article.
NoSQL-database.org used one of my pet hate phrases 'next generation' only ever used as hype with no meaning.
Amusing to think that I first used a 2012 next generation product in about 1970!
Happy Christmas and a great next generation 2013.

Posted: 20th December 2012 | By Leon Guzenda :

IMS is a hierarchical DBMS. IDMS is a CODASYL standard network DBMS. Neither is distributed or fault tolerant. Some ODBMSs, such as Objectivity/DB, which was designed in 1988 to tackle problems that don't perform well on an RDBMS, don't rely on SQL, are distributed, fault tolerant and both vertically and horizontally scalable. So they fit the basic NoSQL requirements.

The fact that NoSQL is still trying to find a standard definition tells us that it's a marketing phenomenon, rather than a single technology.

Posted: 8th January 2013 | By Steven Hall :

A very interesting article that is both 'tongue-in-cheek' and thought provoking.

The NoSQL tag has become quite a misnomer (hence the move towards the Not only SQL definition!) with so many of them offering SQLesk interfaces. Indeed, Mark Logic v6 offers both NoSQL and SQL, and Acunu can utilise its underlying Cassandra's CQL (which has backward compatibility issues!).

That said, the NoSQL market place has certainly been hyped up, but it is no surprise given the technology is still in its infancy (see Gartner Hype Cycle for Emerging Technologies). Its biggest drawbacks generally relate to it not being ACID compliant and subsequently not being able to handle transaction processing (and joins), a lack of user or role level security, and no industry-wide interfacing standard such as SQL.

Somehow I suspect IMS had most of those and that many NoSQL offerings have workarounds! However, whilst IMS pre-dates the internet (and sentiment analysis) could it process high velocity unstructured data from a variety of data sources in real-time? Its all about trying to leverage value and information from confusing amounts of data for competive gain.

Anyway, I suspect that NoSQL will take another year or five to gain acceptance and to overcome most of its inherit and perceived weaknesses. When that happens it will likely end up working alongside traditional technologies - there might even be an agreed name and definition to this new technology by then! In any case, NoSQL (Big Data) is certainly encouraging the move towards cheaper commodity hardware and open source licencing.

The messages above were all contributed by IT-Director.com readers. Whilst we take care to remove any posts deemed inappropriate, we can take no responsibility for these comments. If you would like a comment removed please contact our editorial team.

Post A Comment?

All fields must be completed to submit a comment. Email addresses are passed through to the author so they can contact you directly if needed.