Skip to main content

Are Relational Databases Dead?


I met with a major vendor technology evangelist recently (actually he called himself a technology space regional CTO) on the topic of noSQL and big data. 

Now I’ll admit, like any technologist and architect, I have my areas of specialty and areas of less specific knowledge.  And as the IT industry continues to develop new technologies and approaches yearly, it’s challenging to stay up to date.

So while I know relational databases well, can write good SQL and understand architecturally when to use database power functions like triggers, store procedures and the like, as well as database implementation level issues such as clustering, failover, performance, etc, I had not been paying attention to recent industry changes such as noSQL, graph databases and big data.

This tech evangelist presented the concept of a particular noSQL tool, emphatically stating that the day of the relational database was over.  Relational databases are dead as a future technology.  He even went so far as to predict a major decline in RDBMS use (in new projects) over the next 2 years.

(That’s how you know he’s an evangelist.  Even if a wonderful new tools that makes an older tech completely replaceable with a much better approach arrives, it still takes time to propagate across the industry.)

A recent project brought be face to face with noSQL type challenges.  Businesses are demanding more and more data relationships, more interconnectivity between objects, entities, data elements.  Suddenly businesses expect to build Facebook-like relationships into various parts of their business data, and expect to be able to find and present to users, data by walking the relationship chain.

Lets take an example:

A customer sends an email about a problem with his account.  The business wants to manage the customer relationship, so a link to the email object needs to be placed in the CRM system, as does an entry of what his requests were.  His requests set off a series of business processes to handle them, and each process wants to link to the entity (the original request in the CRM system) that triggered the process.  This allows the process to report “I was started due to…”.  But the inverse is also desired, the ability for the CRM system to look at the request and see the processes started (possibly in other systems) due to the request – so they can report back to the customer (if he calls or looks up the status of his request on a customer web portal) “we did or are doing …… because of your request”.

This bi-directional many-to-many relationship structure would be complicated and tricky to implement in a traditional relational database structure.  (It could be done, but would not maintain any referential integrity.)  But that assumes operating in 1 single database and one single database schema.

What happens when these relationships need to be maintained across modules, across applications and across systems?  When it’s the CRM system and a customer request that’s invoking actions in the Sales system and Billing system?

In other words, what happens when our data relationships are cross platform?

In the past we just threw customer keys at each other.  But when the points of the data relationships move to the tens or more, throwing around keys and indirect relationships is no longer viable.

This is where the noSQL (not only SQL) tools step in.  They are all about building and walking data relationships, and dynamically building data content and data access paths.

So much of our time with relational databases is spent on keys and indices (access paths) and relationships, especially when those relationships or access paths grow beyond 2 or 3.  The recent growth of the JPA Java standard is an attempt to partially resolve this problem (let Java generate it for you).

noSQL graph databases are a full resolution of this problem.

If you haven’t looked into them and you’re into architecture, system integration and/or database design, I suggest it’s time to do so. 

There is one failing though.  Relational databases are very good at what they do, and very well understood for it and it would seem overkill to try to supplant them with noSQL tools in their primary areas of strength.  Yet at the moment there are few access tools that offer a combined RDBMS and Graph-DB approach, where I could build a single query, traverse a graph node path to a relational table row.

Whether RDBMS has a future or not, I don’t know.  But the noSQL and graph database approach is clearly worth looking into for almost any new project.

Popular posts from this blog

Integration Spaghetti™

  I’ve been using the term Integration Spaghetti™ for the past 9 years or so to describe what happens as systems connectivity increases and increases to the point of … unmanageability, indeterminate impact, or just generally a big mess.  A standard line of mine is “moving from spaghetti code to spaghetti connections is not an improvement”. (A standard “point to point connection mess” slide, by enterprise architect Jerry Foster from 2001.) In the past few days I’ve been meeting with a series of IT managers at a large customer and have come up with a revised definition for Integration Spaghetti™ : Integration Spaghetti™ is when the connectivity to/from an application is so complex that everyone is afraid of touching it.  An application with such spaghetti becomes nearly impossible to replace.  Estimates of change impact to the application are frequently wrong by orders of magnitude.  Interruption in the integration functioning are always a major disaster – both in terms of th

Solving Integration Chaos - Past Approaches

A U.S. Fortune 50's systems interconnect map for 1 division, "core systems only". Integration patterns began changing 15 years ago. Several early attempts were made to solve the increasing problem of the widening need for integration… Enterprise Java Beans (J2EE / EJB's) attempted to make independent callable codelets. Coupling was too tight, the technology too platform specific. Remote Method Invocation (Java / RMI) attempted to make anything independently callable, but again was too platform specific and a very tightly coupled protocol. Similarly on the Microsoft side, DCOM & COM+ attempted to make anything independently and remotely callable. However, as with RMI the approach was extremely platform and vendor specific, and very tightly coupled. MQ created a reliable independent messaging paradigm, but the cost and complexity of operation made it prohibitive for most projects and all but the largest of Enterprise IT shops which could devote a focused technology

From Spaghetti Code to Spaghetti Connections

Twenty five years ago my boss handed me the primary billing program and described a series of new features needed. The program was about 4 years old and had been worked on by 5 different programmers. It had an original design model, but between all the modifications, bug fixes, patches and quick new features thrown in, the original design pattern was impossible to discern. Any pattern was impossible to discern. It had become, to quote what’s titled the most common architecture pattern of today, ‘a big ball of mud’. After studying the program for several days, I informed my boss the program was untouchable. The effort to make anything more than a minor adjustment carried such a risk, as the impact could only be guessed at, that it was easier and less risky to rewrite it from scratch. If they had considered the future impact, they never would have let a key program degenerate that way. They would have invested the extra effort to maintain it’s design, document it property, and consider