I think the speed of it, the consistency of it and that it stays up all the time. We've not had any problems with it in the last year. We upgraded to the last version and I'm here taking a look at the newest version that they've released, 1136. We're on 35. It's promised to be a better product, it's much faster and just as reliable. They also have a great web interface that we haven't deployed yet.
Improvements to My Organization
We've had bad systems from other organizations that we've adopted or bought. Workload Automation used to be called AutoSys, and it is actually a better scheduler in my opinion because of the way it schedules. With a base on dependency, events and job triggers. It works on events and triggers. Some of them automatically create jobs and they reschedule them.
AutoSys has it a little differently and it's quite easy to use. It's very easy to set up and it just launches a script anywhere that we have a local agent installed on a server. It goes throughout the world in different locations.
It also works based in Houston, at one of our data centers. We also have some people overseas and we use it abroad. It's a worldwide application that runs over 160,000 jobs throughout our enterprise.
Room for Improvement
I see room for improvement, as far as monitoring the system and having a quiet data center, when you don't have to have people monitoring and watching jobs run or watching flows going and looking for something to stop or a job to fail. I want to be alerted when we have a problem. I don't want to sit and watch a screen or have a staff of people sitting around the world waiting for something to fail. By having a so-called quiet or lights out system, where we get alerted just on these exceptions. That's the direction I'd like to see the product take. You spend a lot of quality time and money on people watching simple things happen. Lights go green, lights go red or lights go yellow. If we only saw something when they went red, those are the kind of alarms and notations I'd like to have to give to a staff of people that can handle those issues and get it restarted.
Use of Solution
The problem has not been with workload, we do have some server outages. Also maintenance times of other products. Workload Automation is dynamic enough to put jobs in pending or put servers offline, until we get ready to bring them on. As soon as they go back online, the server's jobs start rescheduling themselves again. It's a dynamic product, it's been stable and we've never had a real outage with the product.
We have it right now operating in something called dual server mode. If we lose one end of it or one processor, the other side takes over and it picks up from where it left off. It's an always up situation. If we have to throw it back to the primary then we take it down, do an amendments window, do a quick switch over to the primary and let it keep operating.
We never really miss a beat.
As I said, we have agents, our servers, in other locations in different cities and in different countries. We are able to contact with those, schedule batch runs on those and bring the results back to Houston as far as the successes or the failures of those processes.
Customer Service and Technical Support
Technical support from CA has been very good actually. We don't need them very often unless we have a problem with some integration such as a 64-bit application and something that's foreign we're putting on a server, such as BusinessObjects or Oracle, something we haven't seen before. We'll call them for some support. Otherwise our staff is pretty knowledgeable enough and we've had CA products for about 9 years. We're pretty familiar with it on site. It's just when some of the newer products come in, integrating with those, those are the times we've had to call CA support.
As far as the product itself, just learning about some of the new features, we'll speak with their support personnel to find out they operate or how they can implement it with our staff. Once they come on site and given us some information, how-to's, then we pick it up for ourselves. We don't need support as often as we used to with the prior products.
The initial setup, what we had, was called 35 then we went to a 45 version. Now we went from the 45 to what's called R11, that was a nightmare. R11 was a pretty difficult implementation for us. A lot of things changed between the two versions. After we got over those humps, CA put out another service pack and that relieved most of our problems. I think a lot of the rest of the industry suffered some of the same issues that we had.
They were able to quickly release those within 6 or 7 months.
I would give AutoSys a 10/10. Best practices are to plan your workflow. Try to plan where you have as less intervention as you can possibly use. Use the product and the triggers, the timing base events, use the calendars and try to make it flow as smooth as possible. Don't put something that's troublesome into your production environment. Work it out in tests and UAT or development. Even try it in your sandbox if necessary but don't bring it to production.
When it comes to production, if it doesn't work, send it back. You don't want these problems in production. At the shop I work with, we have a 99.91% success rate. When we don't have that, we go through and examine the jobs that fail. If they failed then we have a problem, we examine and get them fixed.
Important buying criteria: reputation, longevity, how is their product and other people's opinions of the product as well. After we've test driven a product, we usually bring something in-house, drive it and see how we like it. If we have use for it, we have enough people that would take a buy in on it, find it's useful, we find it's dependable then we probably want to set something like that in as a candidate. We need to have something that's proven.