A History of Databases in “No-tation”

We’re heading towards very exciting times in the field of databases!

At Topconf in beautiful Tallin, Estonia, Nikita Ivanov (founder and CEO of GridGain Systems) was talking about how the ever crumbling price of DRAM gets in-memory computing and thus in-memory databases within the reach of being affordable by even small and medium enterprises. Nikita claims that 99% of all companies have less than 10TB of transactional data. While this has been completely impossible ten years ago, nowadays, you can store that much data in memory for less than 15000 USD! Compared to the Oracle license that you might buy with the server, that’s almost nothing. Imagine that you can scale up several orders of magnitude without changing your “legacy” architecture. Without switching to something like NoSQL.

A day before, Christoph Engelbert presented Hazelcast, a competitor product of GridGain Systems. Unfortunately, I couldn’t attend his talk but I was lucky enough to spend a couple of hours with Christoph on the flight back home. He’s a very interesting and fun guy to talk to and gave me quite some insight about what his company is evangelising in the context of “Big Data”. Essentially, modern data processing involves moving computation towards data, instead of moving data towards computation. While Hazelcast solves this through their own storage mechanisms, this paradigm has been equally true for “legacy” OLAP systems based on relational databases. Using PL/SQL, or T-SQL, or any other procedural language, you can execute complex algorithms right where the data is: In your database.

For those of you frequently following my blog, you will not be surprised that I am very thrilled about the above evolutions in data computing. The ever increasing consternation with ORMs and the big amount of confusion about the future of “NoSQL” have lead to a recent revival of SQL as a language.

Back to the roots.

This seems to have culminated at the recent O’Reilly Strata Conference, where Mark Madsen, a popular researcher and analyst was walking around with a geeky T-Shirt showing the History of NoSQL. I’ve had a brief chat with him on Twitter. He might be selling this T-Shirt, if it goes viral.

History of NoSQL by Mark Madsen. Picture by Ed Dumbill

History of NoSQL by Mark Madsen. Picture published by Edd Dumbill

So apparently, SQL is back, and strong as ever!

jOOQ Newsletter October 28, 2013

subscribe to the newsletter here

jOOQ 3.2 Licensing Update

Three weeks ago, we have announced and released

jOOQ 3.2 under the terms of a new dual-licensing model offering jOOQ’s support for commercial databases under a commercial jOOQ license.

The jOOQ community has reacted intensively on the user group, most of the feedback having been very positive. The licensing change is received as a good sign for jOOQ’s long-term future, even if some aspects of it were rightfully challenged.

One important such aspect is the jOOQ license 1.0 (Oct 1) phrasing on the term “distribution”. This has been rectified in the new jOOQ license 1.1 (Oct 28) which explicitly grants our licensees the right to distribute and embed jOOQ with their end-user applications (see §6.2). This has been a misunderstanding on our part and was never intended to be a limitation to our customers. As this updated license grants new rights to our customers, it shall be in effect immediately also for existing jOOQ 3.2 customers.

Do also note that we have published an FAQ for licensing questions you may have.

jOOQ 3.3 Outlook

The upcoming release 3.3 (scheduled for early 2014) will include a bulk of exciting new features among which:

  • Support for MS Access, available in the jOOQ Professional Edition
  • Support for keyset paging (see SQL zone, below)
  • SQL transformation SPI example implementations showing how to perform row-level security and shared-schema multi-tenancy with jOOQ 3.2+
  • A new Open Source contribution platform, allowing the community to easily share SPI implementations and other sorts of plugins not maintained by the Data Geekery

And much more. See the GitHub 3.3.0 milestone for an overview of feature candidates, or participate in the user group to discuss your ideas!

Upcoming Events

Some of you may have noticed with regret that the MEDIT Symposium in beautiful Sicilly has been cancelled. It would have been a great place to learn about the latest activities in the Open Source world. But we are more than happy to announce to you that these events are confirmed. Meet Lukas and jOOQ and lots of other good stuff at:

Stay informed about 2014 events on www.jooq.org/news.

SQL Zone – keyset paging

In recent newsletters, we’ve been talking about NoSQL. But let’s get back to SQL andhow to stay in control of your SQL. As mentioned before, jOOQ 3.3 will provide native support for keyset paging, which is also referred to by some as the “seek method”. Our recent blog post explains how this works, and why it allows you to perform constant-time paging even in very large data sets. This is particularly useful when implementing Twitter- or Facebook-like lazy loading of more and more data!

SQL Zone – useful SQL tricks

We also would like to advertise two SQL tricks that come in handy every once in a while:

SQL Zone – MySQL parser / jOOQ code generator

Our October 10 newsletter announced a new cooperation between GUDU Soft and Data Geekery: SQL2jOOQ. Developments around a MySQL -> jOOQ code transformation have progressed quite a bit. This promising tool will soon help you migrate your large legacy application with tons of hand-written, string-based SQL to jOOQ.