Welcome to the January 2018 Continuent Product Management newsletter. It’s the start of the year, and so a good opportunity to look forward, as well as back a little to see how we did. Let’s start with the immediate future first.
- Tungsten Clustering 6.0 is Coming!
- Tungsten Replicator 6.0 is Also Coming!
- Looking at the Year Ahead
- Tungsten Backup, Tungsten Connector
- Tungsten GUI
- End-of-Life Policy
- Release Schedule
- Internal Tweaks
Tungsten Clustering 6.0 is Coming!
The development and restructuring of the product has taken a year to come to fruition, as there are quite a lot of different components, but the new version of Tungsten Clustering 6.0 is due out in February and we’re really pleased with the result.
The focus of this release of the product is to unify the components that previously made up our multimaster clustering solution, that is the new name for what we used to call Multi-Site/Multi-Master (MSMM).
In previous releases, MSMM was built like this:
- Continuent Tungsten on each cluster
- Tungsten Replicator to replicate data between each cluster
Two separate products, two sets of tools and command-lines to handle, and the replicators that handled the cross-site replication were completely independent of the clustering.
In the new multimaster clustering solution, everything is installed as one package. The cross-site replicators are controlled and managed from within cctrl and the rest of the clustering toolkit, and it’s finally possible to get a single unified view of the cluster, state and progress. The connector is properly aware of the relationship, as is cctrl, and recovery and management of the entire process is all within a single environment.
There are some other tweaks and improvements, including the package name, which is now Tungsten Clustering, and not Continuent Tungsten. We’ve got some webinars, deep dives, blog posts and other material to cover all the changes.
Tungsten Replicator 6.0 is Also Coming!
We’ve also made improvements to the replicator, although to be fair it had quite a few improvements and expansions over the past year. We’ve built on that by taking all of that functionality and making some additional improvements, including:
- Re-introduction of the support for writing data into PostgreSQL.
- Significant expansion of the Kafka applier to provide more detailed information in the messages, including whole transaction information so that you can verify message states.
- Support for replicating data from databases that don’t have explicit database primary keys to analytics targets.
There are some other improvements too, so look out for a blog post and release notes to understand everything we’ve added.
Looking at the Year Ahead
Back in October we had an internal meeting where we looked ahead to the next three years of work and focus, and we’re beginning to execute on that.
Here are some highlights of the things we are planning and working on during 2018:
- Two new products (see the more extensive notes on Tungsten Backup and Tungsten Connector below).
- More appliers for the replicator, including new analytics and data warehouse platforms, in-memory databases, and desktop and corporate targets.
- Expansion of our new DDL translation functionality for heterogeneous replication.
- Better tools for schema changes and migration within Tungsten Clustering, particularly for multimaster.
- Better tools to monitor and identify performance and replication issues.
- Easier cloud deployments for some of the most popular cloud environments.
- Faster replicator startup.
- Faster clustering switches and failovers.
- THL compression and encryption at rest.
These are some highlights, and I’m going to keep some items in reserve while we sort everything out.
Something here you would like to see but I haven’t mentioned? Please get in touch with me at firstname.lastname@example.org
Tungsten Backup, Tungsten Connector
Percona Live in Dublin was where we announced two new products that we are working on, Tungsten Backup and Tungsten Connector.
Tungsten Backup is going to be a one-stop solution for handling and automating your backup and restore requirements. It will be based on our existing logging technology, and will use that information to be able to handle:
- Automated backups
- Automatic physical and logical backups
- Backup/restore validation and verification
- Point-in-time recovery
- Per database, table and date data recovery
We’ll also have the ability to restore and extract information and apply back to any target. Want to pull out your year material and apply it into Hadoop? We can do that. Or how about extracting a list of changes to your database and apply into Elastic for some auditing? More information will be coming soon.
Tungsten Connector will be a standalone version of the connector we use within the main clustering product. This means you will gain much of the same functionality, including packet inspection, query distribution, load balancing, but you will be able to use it with other database deployments.
In fact, you’ll be able to use it with a variety of databases, so if you want to place the connector in front of your Oracle databases, or PostgreSQL, or even a Vertica cluster, you can do so. We’ll have more use cases, deployment examples, and capabilities to share soon, so look out for the blog posts and webinars that will surely be coming down the line.
We have a surprise to come, in that we’ve been developing a GUI that enables full control of your Tungsten Clustering deployments, all through a web browser, while also providing a single view of your cluster state and performance.
Having worked on it for a few months, we are surprisingly close to a release, although from a practical level it’s probably going be the beginning of the 2nd quarter this year when we finally make it available.
We are, however, very pleased with the result and right now it’s going through a combination of UI development, user testing, security and QA, to ensure that the product is ready to be in the hands of our customers.
A GUI is something we’ve wished for here at Continuent for many years, and we finally have one that provides the ideal combination of functionality and information. Believe me when I say the excitement in our respective distributed offices is intense!
Back in November last year we made the decision to have a more formal end-of-life policy for our existing products.
There’s a longer more detailed description that will be going out, but the moment the basic approach is for Continuent to support their products for three years from the date of the launch of the original major version. So, for example, 6.0 will be coming out soon, and we’ll support that right round until February 2021. All 6.x versions will have the same EOL date.
This goes hand-in-hand with our new release schedule, which we’ve actually had in place for the last 18 months or so now. That stipulates that we have a new major release every year. So by the time 6.x support ends in 2021, we will already have released v7 in 2019, v8 in 2020 and v9 in 2021.
Look out for a more detailed description of the policy soon.
As mentioned above, we changed to a more aggressive release schedule in 2016 so that we could ensure that we get new features and functionality, and bug fixes, in to your hands as quickly as possible. The official policy is:
- One major release version every 12 months (e.g. v6.x, v7.x)
- One minor (feature) release version every 3 months (e.g. v6.1, v6.2)
- One bug fix release every 6 weeks (e.g. v6.1.1)
How did we do in 2017 on those goals? Here’s a list of the releases:
5.0.1 23 February 2017
5.1.0 26 April 2017
5.1.1 23 May 2017
5.2.0 19 July 2017
5.2.1 21 September 2017
5.2.2 22 October 2017
5.3.0 12 December 2017
Actually not too bad. Towards the end of the year the releases started to spread out a bit, but we still had a release of some kind almost every two months or less.
As anybody who has worked in software development will tell you, sometimes you have stop, take stock of what you have and where you are, and determine how you are going to make and produce everything that you’ve promised over the coming months.
We decided back in a meeting almost exactly a year ago that we need to make some changes to our internal systems. We have, for example, been using Jenkins as our build and test environment for a while, but we’ve begun to outgrow it, especially in light of new products and the new release schedule.
So, once our 6.0 release goes out the door we are going to be doing some housekeeping. We’re migrating to a new QA environment, we’re going to be making some changes to some of our internal processes, and we’re going to work hard to ensure that the features and functionality we’ve got scheduled over the next few years will get ordered properly.
That does mean we’re going to be quieter during February and March than usual, but it will all be for a good cause in the long run.
If you have any questions from the newsletter, please feel free to get in touch with me directly (email@example.com) or use the usual firstname.lastname@example.org or email@example.com addresses and we’ll do our best to answer your questions (secret plans and hidden projects notwithstanding!).