IBM recently (April 12th 2016) announced the latest version of the DB2 for Linux, UNIX and Windows. The worst kept secret in the world, that “vNext” would be V11.1, was finally out in the open. By the time this article is published the GA date should be just around the corner (the planned GA date in the announcement is June 15th 2016).
Over the course of the next month, the IDUG Content Committee and contributors from IBM and third-party experts will bring you detailed information about the new release. The information will be provided by those who either developed the code or have been working with it during the Early Access Program which IBM has been running for some time.
This introductory article will talk in broad terms about the highlights of the release, and then will go into more detail on changes to packaging and the delivery and support model. Subsequent articles will then go into a lot more detail about the various technical enhancements.
DB2 for LUW V11.1 Highlights
In this section we will cover the highlights of the new release. Some of the highlights will be the “big ticket items” which have been widely publicized by IBM during the announcement (and also covered in detail at the recent IDUG North America DB2 Technical Conference 2016 in Austin). But we will also cover some less well publicized, but we feel important, new features.
The biggest features in this release relate to enhancements in pureScale and BLU functionality. However, there are significant improvements in other areas as well.
For pureScale there has been a focus on ease of deployment and operation. Previously setting up a pureScale environment involved many steps, many of them dealing with complex networking and GPFS file system commands. In the same way that db2haicu was introduced to streamline the setup of the TSA/MP clustering for HADR environments, the db2cluster command has been introduced to wrap much of the complexity of configuring and managing pureScale environments, thus making pureScale more approachable for non-expert installers. In addition to an easier setup, pureScale has been enhanced to support operation in HADR SYNC and NEARSYNC modes, thus providing more robust remote disaster recovery capabilities in combination with the local high availability provided by pureScale.
On the BLU front the biggest single feature is support for BLU in DPF environments, thus expanding the scope of BLU to the largest analytics environments. At the same time IBM continues to lift the restrictions on SQL processing that can be performed completely in the columnar engine, as well as introducing improvements in performance of updating statements (insert, update and delete).
There are two major security enhancements in this release. The biggest of these is the ability to use enterprise key management systems for storage of native encryption master keys, eliminating the need to have a local key store. A less well publicized, but equally important, feature is the ability to encrypt the HADR replication stream using TLS
Following feedback from major customers there has been some significant improvements in the upgrade process. To help those who have slipped behind on upgrades, it will be possible to upgrade directly to DB2 V11.1 from as far back as DB2 V9.7 (this is likely to be a “once only” opportunity to catch up by three releases in one step). Those who have been keeping up to date with maintenance, upgrading to DB2 V11.1 from DB2 10.5 Fixpack 7 or higher on single partition and pureScale environments will no longer be required to take a full offline backup before and after the upgrade, since a process now exists which will allow rollforward through an upgrade. This feature needs close attention however, as there are limitations as to what it allows (watch out for more on this in the weeks ahead).
A lot of work has been done to enable DB2 for operation in hybrid (mixture of cloud and traditional on premise) deployment environments. These improvements have been put in place in support of IBM’s own cloud based offerings such as “dashDB” (now available in both “BLU” and “transactional” flavours) and “DB2 on Cloud”. The first fruits of this work appeared, largely unheralded, in DB2 V10.5 Fixpack 5, with the introduction of the db2RemStgMgr command. This was a command line tool which could be used to store and retrieve DB2 backups from cloud storage solutions (initially Amazon AWS S3 and SoftLayer ObjectStore). DB2 V11.1 builds on this to build support for directly reading / writing DB2 backups, restores, loads, exports and ingests directly from the same cloud storage providers. In many ways this starts the process of utilizing cloud storage solutions in a similar way to TSM has been used by many DB2 customers for many years. However, the functionality is not yet complete, since archive logging cannot yet be done directly to cloud storage.
Finally, no summary of DB2 11.1 features would be complete without mention of dsmtop. This is a replacement for the much-loved db2top tool, which has not been enhanced since DB2 V9.7. It not only provides much of the same functionality as db2top, but also fully supports new features such as pureScale and BLU. It uses the latest DB2 monitoring functions and is a welcome addition to the toolset (and an excellent example of how customer feedback is listened to by IBM).
Licensing and Support Model Changes
As with all DB2 releases there are changes to the way that the product is licensed. In DB2 V11.1 there are two major changes. Firstly, DB2 Express Edition is no more: however, pricing of Workgroup edition has been significantly changed to make it still accessible to entry level users. DB2 Express-C (free and unsupported version) will still be provided, but for those who want support then Workgroup Edition is the entry point. That is unless you opt for the second major change to the licensing model. It is now possible to lease DB2 for a monthly license charge by choosing the DB2 “Direct” editions. These come in two functionality levels: DB2 Direct Standard Edition (roughly equivalent to DB2 Workgroup Edition in functionality and limits) and DB2 Direct Advanced Edition (roughly equivalent to DB2 Advanced Enterprise Server Edition, in that it includes all functionality and no limits on capacity). In conjunction with these new “Direct” editions a new charging unit has been introduced. This is called the Virtual Processor Core (VPC), and is the price per month for using DB2 on one virtual (or in bare metal implementation physical) processor cores. The name reflects the fact that this is a purchase metric designed for cloud based or hybrid deployment scenarios.
One other licensing change which is worth mentioning is the “DB2 for Big Data” licensing option. This allows you to use your license entitlements to deploy either DB2 or BigInsights (IBM’s packaging of Hadoop), and to move licenses between the two as your needs change.
Perhaps more significant than the licensing changes are the proposals for the DB2 support model going forward. Currently IBM delivers a single stream of maintenance, with Fixpacks (which can include a mixture of fixes and new functionality appearing every three to six months). This does not fit the needs of all customers. Some would like more rapid availability of new functionality, comparable to the rate of new feature deployment appearing in IBM’s cloud based offerings (particularly dashDB). Others would prefer to only receive fixes, with no new features, except for well-defined and widely spaced release boundaries.
IBM is therefore proposing to move to a dual paced DB2 delivery model for “on premise” DB2. This is expected to start following the delivery of DB2 V11.1 Fixpack 1 (Fixpack 1 traditionally is a “sweep up” for features which were in the roadmap for the release but for one reason or another did not make the GA build). After that point there will be a “Feature” stream and a “Fixes Only” stream. Once a system embarks on one or the other stream, crossover to the other stream will not be possible until the next “release boundary” (perhaps 2 to 3 years hence). So users should think very carefully about what their plans are for the next few years. It is obvious, both from monitoring the “cloud first” deployments of functionality based on DB2 and in involvement in the DB2 TAB (Technical Advisory Board) that IBM plans to deliver significant new features in the feature stream. Do you want to have access to that functionality as soon as it is available? If so, you want to choose the “feature stream” (or maybe consult with IBM how you can be involved in the DB2 TAB process, so that you can have early access to information on planned features so that you can make a more informed decision). Or is your organisation conservative, and does not wish to exploit this new functionality for a considerable time? Then the “fixes only stream” may be for you.
DB2 V11.1 pushes DB2 forward in a number of significant areas. There is still more work to be done, particularly in rounding out the availability of all DB2 functionality across the pureScale and BLU offerings, but some major steps forward have been taken in this release.
This opening article only gives a brief overview of some of the key new features. Continue to follow the IDUG Content blogs for more information in the weeks ahead.