Ajilius in 2018

As 2017 draws to a successful close, let’s take a few moments to look at some up-coming changes to Ajilius.

Licensing

You’ll notice from our Pricing page that Ajilius is splitting into two Editions – Standard and Enterprise. Standard is for smaller, SMP data warehouses. Enterprise is for large data organisations wanting the ultimate in scale and performance.

This will provide a lower-cost entry point for small-to-medium organisations, for whom the full benefits of Enterprise Edition may not be realised.

All existing users will be automatically converted to Enterprise licences. As usual, you will pay no more than your initial licence price.

All existing opportunities will reflect our current pricing and licensing. Only opportunities commencing after January 1, 2018, will be affected by the changes.

Features

We’ve also got some major changes coming in V4, to be delivered by phase during 2018. Here’s a brief summary of the highlights:

  • Data Vault 2.0 support, for both Standard and Enterprise Edition.
  • Data Lake support, for Enterprise Edition.
  • Star Schemas built from source, vault or lake.
  • Data Vaults built from source or lake.
  • Data pseudonymisation on ingestion
  • Streaming ingestion
  • Forward chaining to drive near real time fact tables
  • Major enhancements to metadata store, including versioning and copying/cloning.
  • Containerisation, with full support for Docker deployments
  • Support for OpenJDK/JRE as well as Oracle JDK/JRE

Please contact us if you’d like further details of any of these changes.

Ajilius. Striving to be the best in data warehouse automation.

Ajilius Release Numbers

We recently finished an RFP submission with a question that asked:

Describe your version numbering system, and how well your product keeps to it

I thought the answer might be useful to all customers.

Our version numbering system

The Ajilius release number is comprised of three elements – major version, minor version, and point version. For example, in version 3.2.1, the major version is 3, the minor version is 2, and the point version is 1. But what do the various types of version mean?

A major version represents an architectural change. For example, Python to Java was a major version. the CTAS engine was a major version.

A minor version represents a metadata change, or a module-level change. Metadata changes are where we alter the structure of the metadata repositories used by Ajilius. Module changes are where we introduce major blocks of functionality. Incremental deployments were a minor version, SSAS integration was a major version.

Point versions are where we make relatively minor changes to the application code. These may be error corrections, or enhancements that build on existing functions or metadata.

How well do we keep it?

When I look back over the Mercurial logs for Ajilius, I’d have to say that we’ve only been 60-70% compliant in our numbering scheme. We’re always correct with Major versions, always correct with minor versions, but we’ve often let small metadata changes slip through in point versions.

From now on, though, we’re going to be 100% compliant with this structure.

Why bother? Simpler upgrades.

Better adherence to our release structure will help us to simplify our upgrades. We’ll make it faster and easier for you to download and apply new code.

  • We’ll always offer a full installer. This will create a completely new installation.
  • We’ll always offer an upgrade installer. This will upgrade any earlier release to the current release, including metadata structure changes and library dependency updates.
  • Now, we’re going to simplify point releases, as just the Ajilius executable program. Point releases will comprise only ajilius.jar program, currently around 400k in size. This will be faster to download, and more simple to copy to the Ajilius folder.

So when you get release 3.3.2 later this week, you’ll know that it is a code-only change to Release 3.3, and makes no changes to metadata or libraries.

Ajilius. The best development cycle in data warehouse automation.