We have been experimenting with centralized connection profiles. There are some bugs to be worked out. So, we don't feel 100% comfortable with only using centralized connection profiles. We do have a mix of control on agents out there, which leads to some complications because earlier agents do not support centralized connection profiles. A lot of the areas of improvement revolve around Automation API because that area is constantly evolving. It is constantly changing, and it is constantly being updated. There are some bugs that are introduced from one version to the next. So, the regression testing doesn't seem to capture some of the bugs that have been fixed in prior versions, and those bugs are then reintroduced in later versions. One particular example is that we were trying to use the Automation API to fetch a number of run ads users from the environment. The username had special characters and backspace characters because it was a Windows User ID. In the documentation, there is a documented workaround for that. However, that relied on two particular settings in the Tomcat web server. I later found out that these settings work out-of-the-box for version 9.0.19, but those two options were not included in the config file for 9.0.20. So, it led to a little bit of confusion and a lot of time trying to diagnose, both with support and the BMC community, what is the issue. Ultimately, we did resolve that, but that is time spent that really shouldn't have been spent. It had obviously been working in 9.0.19, and I don't know why that was missed in 9.0.20, but that's a primary example of an improvement that can happen. We've also noticed that the Control-M agents themselves now run Java components. Over time, they tend to destabilize. It could be because garbage collection isn't happening, or something is not happening. We then realize that the agent is consuming quite a large amount of memory resources on the servers themselves. After recycling the agents and releasing that memory, things go back to normal, but there are times when the agent becomes unresponsive. The jobs get submitted, and nothing executes, but we don't know about it until somebody says, "Hey, but my job isn't running." When we look at it, it says Executing within the GUI, but there is no actual process running on the server. So, there is some disconnect there. There is no alerting function or the agent there that says, "Hey, I'm not responding." It is not showing up in the x alerts or anything like that. The integrated guides have not been that helpful to us. I do find a lot of the how-to videos on the knowledge portal to be useful. However, there are some videos where the directions don't always match with some of the implementations. There are some typos here and there, but overall, those have been more helpful for us. Its pricing and licensing could be a little bit better. The regular Managed File Transfer piece, is a little overpriced, especially for folks who already have licensed Advanced File Transfer. What I'm also noticing when I'm trying to recruit for Control-M positions is that the talent pool is quite small. There's not a whole lot of companies that utilize Control-M, and if they do, most people don't want to let their Control-M resources go if they're good. There is a high barrier of entry for most people to learn Control-M. There are Workbench, Automation API, and so forth mainly for developers to learn, but there are not a whole lot of resources out there for people to get more familiar with administering Control-M or things like that in terms of the technology or even awareness. So, it becomes very challenging to acquire new resources for that. A lot of the newer people coming out of college don't even know what is Control-M. If they do, they think of it as a batch scheduler, which is certainly not true in its current transformation. Control-M is a very powerful enterprise tool, but the overall perception has not changed in the past five to six years that I've been working with Control-M. There's not much incentive for people to dive into that world. It is a very small community, and overall, the value of Control-M is not being showcased adequately, maybe at the C-level for corporations. I've had multiple conversations with other people and other companies who have already exit using Control-M. About 70% of the companies out there do not take full advantage of the capabilities in Control-M. So, that type of utilization really hampers and hinders the reputation of Control-M. That's because people then acquire this untrue concept that Control-M can only do X, Y, and Z, rather than the fact that Control-M can do so much more. I don't know if it needs a grassroots marketing movement or a top-down marketing movement, but this is what the perception is because that's what I'm hearing and that's what I'm seeing. For some of the challenges that I face working in Control-M, when I go back to my management and say, "Hey, I want to spend more money in this space," they're like, "Why? Can you justify it? This is what we see Control-M as it is. It's not going to bring us value in this area or that area." I have to go back and develop a new business case to say, "Hey, we need to upgrade to MFT enterprise or something like that." So, it definitely requires a lot more work convincing management in order to get all these components. In the past, we had to justify acquiring a workload change manager. We had to justify acquiring the workload archive. All of these bring benefits not only to our audit environment but also to the development environment, but the fact that we had to fight so hard to acquire these is challenging.
