What is our primary use case?
We use it for promoting and demoting to match our SDLC. So, things start off in development, we build it – whatever it is – and it moves to a testing region. We're trying to shift that left as part of our future state where it's automated regression. But right now it's, especially on the mainframe, it's mostly manual. We move it up to testing. If they're happy with it, it goes to integrated testing, which includes customer acceptance testing. After that, it's moved to QA where it is run for a certain period of time, depending upon different application requirements in an environment that very closely mimics production as close as we can get it. And then when it signed off there, ChangeMan handles everything except the actual move through the hole in the firewall.
Why it doesn't do that? I don't know. To me, it seems like it should be able to. We put it onto what we call an MFT server, which moves it to another server that's inside the DMZ for the mainframe application.
On the distributed stuff, we run four different anti-virus suites against it, looking for vulnerabilities, penetrations, things like that. On the mainframe, I don't know that we do anything other than have a code to review with the Ops folks.
What is most valuable?
I liked the fact that the skeleton design is very flexible and allows for concurrent development. We're currently talking to Serena ChangeMan about how to get to parallel development and whether or not they support it in a Git-style with branching.
What needs improvement?
I would like to see them enable parallel development for online. It's available now for batch stuff on the mainframe. Jenkins, IBM, and Rocket all supposedly already have safe and workable version of Git for the mainframe. With that in mind, we need to know where our feature is. Unfortunately, as good as ChangeMan is, if we're listing downfalls, they should be talking to me about this. Not me having to go talk to them about it. Hopefully, they have this in the works and they are positioning their product for the future. That's just a straightforward comment. It really comes down to whether they are complacent or not.
The other thing that should improve is cost. Git, even on the mainframe, a large part of the components "free". They're the way you set them up yourselves and they're self-sustaining and built into the infrastructure. You can't charge me for that.
I gave ChangeMan an overall score of eight, but what could drive them to a nine is coming up with a way for all my code that belongs together to be tracked and moved together. Everything should move based upon the mainframe; it needs to be alerted and synchronized.
For how long have I used the solution?
I have been using this solution for about 16 years.
What do I think about the stability of the solution?
It's a main prime app, so it's very stable. Would I say it exceeds my expectations? No, but that's because mainframes are a single point of failure; they have to be stable. And we've had no major outages that I'm aware of with the product. I never had any loss. We'd never lost code. That's important for source control management.
What do I think about the scalability of the solution?
Scalability is great. It has absolutely met every need for us so far. We do have some concurrent development paths and we're able to flexibly assign variables. At the same time, our skeletons assemble where we want them to, so the scalability is very good.
How are customer service and technical support?
I've never had any problems with their support, whether it was simple questions or complex questions. They've been Johnny-on-the-spot with the right answers.
How was the initial setup?
ChangeMan was very complex to set up. Its complexity is based around the fact that it's more flexible. It lets you do a little bit more than its competitors in slightly better ways and that's a good thing.
At my previous job, it took us about 18 months to fully deploy ChangeMan from A to Z.
What about the implementation team?
At my previous job, I was part of a team that deployed ChangeMan in-house. I highly recommend doing it yourself if you're going to be supporting it going forward. We did not have integrators because we wanted t come out the other end knowing ChangeMan fully.
What other advice do I have?
My advice is: Take the time to learn it. It's very flexible and it will do what you want, so take the time to learn it.
Which deployment model are you using for this solution?