IBM MQ Review

For me, the most important features are its interfaces with RACF and how we can set up the security to allow and disallow who can get to it, who can use it; and then what MQ gets to do on the mainframe and what it does not get to do, basically.

Our organization uses it a lot to interface applications that are outside the mainframe with applications on the mainframe, or to CICS, items like that.

It helps meet that threshold between what do the application people want to do – because they want to do everything now on GUIs and outside applications – and be able to have the security of the data living on the mainframe and how they get to it. It's the go-between between those two worlds.

There are probably dozens of ways we are using MQ to better connect across cloud, mobile, and devices, but it's mostly the fact that they are setting up stuff and then they use the MQ as the go-between between the distributed world and the mainframe. That's mostly what it's being used for.

Sometimes the applications people don't really understand MQ. For example, we had somebody set up a call through MQ and they ended up making dozens and dozens of calls when they only really only needed to do one. They don't understand how MQ really works, and how it pulls the data and then distributes it back to them, etc.

I think the application people understand that MQ can do it, but they don't really understand the mechanics behind it. They need to be better educated; how to use MQ, get the data that they need, and not cause conflicts.

At the level of the application development people, there needs to be more communication, more information that they have so they understand, because, in essence, what you're using MQ to do is to go to the mainframe and get things. They're so used to their Windows environment, and they don't really understand how MQ grabs that data, and what the mechanics are behind the scenes. And I think that the applications people need to better understand it. Or else something put into MQ so that it is more obvious to them. They don't know what to ask for. They just know, "We're going to go against this data" and they don't know the difference between the different types of security they can set up. The different access and the different classes. We use different classes in RACF; they have no clue what a class is.

There either needs to be better education on there, and or some tools built into MQ that helps them know what to ask for.

I have a very high impression of the stability of MQ; we haven't had any problems with it. MQ has been very stable. I think we've had it go down once since I've been here, but it was due to something somebody screwed up somewhere else, not MQ's fault.

So far, we haven't had any scalability problems either, but we're only about a year and a half into this.

I have not had to use technical support. I've had to use IBM technical support because of some issues, but I never had to talk to the MQ people. We have an MQ rep on site and he handles that stuff.

I was involved in the decision process of how we were going to use RACF, and what they were going to set up to do their calls, but they decided they were going to use MQ. I was actually called in as a RACF specialist to help get that interface going.

Before you implement it into RACF, really investigate the classes and how you're going to set those up, and make sure it's clear with the application development folks. Especially if you're trying to test QA and production separately, it's really important how those classes are set up, and how you set up the instructions for those guys.

The most important criteria for me when selecting a vendor to work with are stability, technical support, obviously the more customers they have in a similar type of field; that's probably what's most important to us, generally.

So far, we've had good luck with it. It seems to be working and it seems to be very stable.