Introducing - Connections Pink
With the landscape for collaborative technology becoming increasingly competitive, IBM are rebuilding their established Connections solution from the ground up, making it more scalable, flexible and prepared for the needs of the modern customer.
Ben Williams, Collaboration Engineer at Choose Portal, discusses how he got this highly customisable open platform working for him, drafting in the help of some fellow IBM partners and IBM’ers to truly put this revolution in software delivery to the test.
What is Connections Pink?
Connections Pink is far more than just a rebrand and a change in colour scheme, it’s a huge shift for IBM software. Pink is a complete code rewrite which will be built on open source foundations – moving IBM away from building software for customers and toward building software with customers.
While IBM Connections has been a fantastic platform for collaboration, with plenty of options and some open API interfaces to provide a level of customisation, you were still tied to IBMs processes – including lengthy release cycles.
Not with Pink.
With virtually everything being API enabled, Connections Pink is ready for continuous updates and integration with third-party apps. The code base stays the same, either Cloud or On-Premise but you gain control of where your data and components live - Cloud, On-Premise, or a mixture. It’s up to you what lives where.
Moving to Pink
Orient Me is the first part of the Pink "infrastructure" in which we will see technologies such as Kubernetes (K8S), Docker, MongoDB, Redis, NGINX, Node.js replace the traditional IBM technologies of WebSphere, DB2, TDI (SDI) and IHS.
The most satisfying aspect of getting Orient Me up and running is not the fact that it works but the journey I had to take to get there. The approach to this transition will be phased as detailed in the following image.
As we’ve mentioned previously, Connections Pink goes beyond a rebrand, it’s a complete overhaul which brings in a new set of technologies. While this is an exciting development, it poses challenges for those that have administered Connections and are used to IBM technology.
One of the challenges I personally faced when getting Orient Me working on my lab servers were errors relating to MongoDB within various K8S pods and even the MongoDB pods themselves. I spent a great deal of time trying to work out the problem. I even wrote about it about on my blog here.
A follower and IBM’er, David McDonagh, who was referred to my blog by another IBM’er Eamon Muldoon, picked up on my problem. He was very helpful, passed on some suggestions and we exchanged messages. David’s preference was to respond via my blog so that others who might be experiencing similar issues could see our exchange and potentially find it useful. It’s great to see an IBM'er wanting to share our efforts with the wider community.
Our exchange did eventually dry up and I found myself becoming increasingly frustrated in the run up to a business partners event in Dublin on Pink which I attended remotely. IBM laid on Bluemix servers for all attendees (in Dublin and remote) to install the technology I mentioned earlier to get Orient Me working. To my frustration, it did. During the event IBM provided us with tasks to customise Muse, demonstrating what is possible. The three days were a huge success, during which both David and Eamon were present to discuss the software with attendees, providing reassurance I’d been talking through my configurations issues with the right people. The week after I was lucky enough to visit Dublin with a customer to talk about XPages, Pink and Watson Work Services/Workspace. I used it as an opportunity to get in touch with David. Eamon, who was our host for the two days arranged for David and a colleague who worked on MongoDB to come down for a couple of hours to look at my servers. If this was a localised problem in which I was the only person affected I wouldn't have pressed so hard but I was aware of at least two others who were experiencing the same issues and errors.
David and his colleague Bruno worked on my servers, checking logs, running through the same steps I had done half a dozen times prior, but they noticed that the servers were running for a long time with CPU constraints. Running my local servers on a large VMware ESXi server I expected some resource constraints because these servers are test servers and do not have the resources available to them unlike production servers. After a while they both told me to give my servers more resources as that was the only difference between my test servers and those provisioned in Bluemix. After getting home I added more resources to my server, lo and behold the errors I saw previously were no more.
Now this may not sound like a surprising story in its own right but these are the things that have left an impression on me, which leads me to believe that Pink will be a success. Here’s why:
People such as David who are enthusiastic about the journey Connections is taking. People who are enthusiastic about the technology. People who are willing to take time out of their busy day to speak with me at the Dublin labs and help fix my problem. People who are willing to comment, publicly, on blogs. People who join the Connections Skype chat to help customers and business partners with their problems outside of the formal PMR route. It seems that this approach and willingness for Pink to succeed has proliferated throughout the teams working on Pink, spearheaded by Jason Roy.
The openness of the technology is attractive to many people. How many people can get started with MongoDB and how many of those people are able to do the same with DB2, Oracle or SQL Server? Speaking with a customer earlier this week they remarked that they are using some of the technologies internally, starting to use containers for their developers and are interested in knowing more. IBM are using containers, Solr, Elasticsearch, Kibana internally. It seems that IBM is starting to replace some of the traditional IBM tools with these to ship, index and visualise logs.
Moving to a continuous delivery type deployment model using agile methodologies is on-trend at present but it makes a lot of sense. It may seem more practical for larger organisations to do this but some of the subtleties can be used by smaller companies too. For those companies that use Connections what will this approach mean? IBM will follow the continuous delivery approach and as consumers of the software we will need to be aware that there will be many more releases.
The release process
For on premises customers, previously, a new version or even cumulative release would mean a great deal of work. IBM has proposed that with Pink, the release process will be much simpler. A new release will be available from the "mothership" aka IBM's GitHub or similar which will be pulled down to the on premises servers running the Pink containers. How this will happen is still being decided but the aim will be to use existing approaches such as pulling from GitHub or in the short term downloading a zip and deploying the application locally.
It is these reasons why I believe Pink will be a success.
If you’re interested in learning more about the IBM Pink and the deployment model, get in touch with one of our expert team at Portal. Alternatively, join the conversation over on our Twitter page: @chooseportal