What is our primary use case?
We use it for IT workflow automation. We have three main categories: billing, reporting, and analytics. We automate the entire workflow, end-to-end, including the file movement from our managed file-transfer solution to our NAS storage point and then kick off the downstream workflows. Most of them are ETL-based in Informatica. And the loads are done to a database and/or storage with output. It's the entire workflow.
We host it on-prem.
It is our enterprise scheduler; we're migrating everything to it. We'll only have Stonebranch as the enterprise scheduler controlling all of our jobs when we get to the end of our project. So it's being used extensively. We'll be somewhere north of 450 jobs when have completed the project and we will increase usage. As more workflows come online, Stonebranch is the platform we will use to automate them.
How has it helped my organization?
Stonebranch Universal Task allows us to break up our workflows into tasks that are singular and focused and we're then able to string them together by drag-and-drop. What we did is we built up a group of tasks that are templates, so to speak. Once those task templates were built up, it allowed us to go in and build workflows relatively quickly, because all we had to do was fill in the particulars of what we needed that task to do, without building the task from scratch.
One of the biggest issues with workflows is that the completion of a step. When a process ends, it needs to generate a completion through Standard Error or Standard Out, or it needs to generate a completion file so that the workflow knows that the step is done and that it can kick off the next workflow task. We've been able to impress upon our ETL developers that, when they're done, they need the process to drop a completion flag so that the workflow knows it's done and ran successfully. They really didn't deal with that before, because they were using disparate schedulers in disparate systems and simply setting the clock: The file gets here at two o'clock, we'll kick this job off in this scheduler at two-thirty. And then the other technology, downstream in the workflow, we'll set that one for three o'clock. But you end up with problems if the previous workflow step isn't complete when the next step kicks off. The job fails and blows up.
Stonebranch eliminated those dependency gaps because we know for sure that the previous step was completed before we kick off the next step in the workflow. We've eliminated the vast majority of that kind of problem.
Stonebranch has saved us money. From an operational standpoint, we're in a position now where we know when workflows fail and we're not missing SLAs as much. Very infrequently do we now miss an SLA because, if a workflow fails, we know it failed. That means the support teams have the ability to go in and address it before we miss an SLA. That's been the major cost savings.
Had we stayed on Cisco Tidal, to do the caliber of work that we've done in Stonebranch, we would have had to hire additional people. I only had one person on Tidal and those resources would cost more than the resources I'm using now, because of the specialized skillsets you need to do the work in Cisco Tidal. On our automation team, we have one senior person. We took two guys out of our network operation center and put them with him. They're Linux administrators and they learned the Stonebranch product over the first year and got very proficient in it. We're doing automation but their cost to the company is a lot less than going out and hiring somebody with the skills in Cisco Tidal or Automic or one of those other platforms. So there is an HR resource savings as well. I would estimate we're saving about $20,000 a year per head-count.
Stonebranch has also saved us man-hours during implementation and production. If I compare it to Cisco Tidal, to build a workflow in Stonebranch, beginning to end, is less time-consuming and less complicated than it was in Tidal. I'm sure it has saved us money. It has saved us about 20 to 25 percent per workflow. When you spread that over 400 workflows, it's a pretty significant number.
What is most valuable?
The Universal Agent is the most valuable feature. Being agent-based and being able to go across multiple technology stacks, which is what our workflows do, Stonebranch gives us the ability to bridge those disparate technologies. It enables us to remove the dependency-gap with the agent so we know the status of the workflow at each step.
Workflow development in Stonebranch is straightforward. There is something of a learning curve, but it's not very steep. Being able to develop workflows without having to train and develop some very specialized skillsets to use the tool is very useful.
Stonebranch absolutely helped enable digital transformation in our company and it still is. In our automation efforts, we're pushing everything to Informatica and, as we move those ETLs, we're automating the entire workflows. In phase-one and phase-two, there were 244 jobs migrated in from other ETL platforms to Informatica, and we've automated all of those. We have almost 200 jobs remaining. We're going to have something approaching 450 workflows in Stonebranch when we're done.
For how long have I used the solution?
We have been using Stonebranch for just under three years.
What do I think about the stability of the solution?
The Universal Controller is very capable, very stable. We really haven't had any problems, stability-wise, with the Controllers. We've been very happy with the resiliency of this solution. We really don't have problems with Stonebranch. Once we get agents in place and configured, they're stable. Our Stonebranch Universal Controller runs. It is very stable.
Occasionally, we have an agent that doesn't come back up after OS patching. That doesn't happen very often. But everything is getting patched on a monthly cycle on the analytics side when the repos come out and get pushed out from satellite. But that issue is so infrequent that we don't make a whole lot of noise about. That's probably the most common issue, though: After patching, every once in a while we have an agent that will hang on the way back up. It's really just a restart of the agent and it comes back up. But that might be one thing that could be improved.
What do I think about the scalability of the solution?
We're yask-based, so if we get to a point where we need to run more monthly task, we can scale our license to monthly task.
We haven't come anywhere close to having performance problems or capacity problems running the jobs that we're running at this point. But if we needed to scale larger, because we are transaction-based, it's a matter of scaling up or scaling out the Stonebranch Controller. With the 450 jobs that we're going to be running, we're running on a single Universal Controller. We have DR, we have it in another data center as well, the databases are replicated. But with the one Universal Controller on a good size virtual machine, and being transaction-based, being able to run 400 to 500 workflows is good. We can scale horizontally or vertically as we need.
How are customer service and technical support?
When we have a problem, we create a support ticket with Stonebranch. Usually, inside of two or three hours, we're talking with somebody at Stonebranch. They're really responsive and we're very happy with the support side of the product.
Which solution did I use previously and why did I switch?
We had Cisco Tidal going before Stonebranch. It was an older product which was being retired. Cisco's core competency is not automation, it's networking. The product was somewhat clunky and difficult to work with. Stonebranch is a lot easier to work with. We made the decision to do a PoC and find a different scheduler and selected Stonebranch.
How was the initial setup?
The initial setup was very straightforward. Everything is running on virtual machines, such as the installation of the Controller and the installation of agents across the first several platforms which we set up to do the integrations on with Universal Agent.
Start to finish, until we were ready and configured to start building workflows, the deployment took inside of about four days.
Our deployment plan was building a Stonebranch environment and connecting to our current ETL environment and storage environments, so we could start automating ETL workflows. We had a couple of problems with the NAS storage. We haven't had that problem since. A lot of it was just because of the particulars of how our NAS was set up. That was the only complication we ran into.
We were able to get deployed and configured according to our plan, to set up the Universal Controller and agents to get across to our ETL environment and our storage environment. Within those four days, we had built our first workflow with the first job which we had targeted for the implementation work with the Stonebranch.
What about the implementation team?
The Stonebranch team was onsite when we deployed, so the installation was pretty straightforward. We had support. They were outstanding. The two gentlemen who worked with us through our PoC and install - Tracy and Mark - we still talk with those guys quite often. We're part of the user community and we get up to Atlanta for the Stonebranch conference every year. They're at our building about once a quarter.
What's my experience with pricing, setup cost, and licensing?
We're transaction-based, as far as our licensing goes. We have 50,000 task a month and our licensing cost is $55,000 a year, if I remember correctly. There are no additional costs. Support is included in that.
At the time, Stonebranch was about 30 to 35 percent of the cost of Automic. Because we also run IBM Sterling File Gateway, and we run IBM WebSphere and IBM middleware for our applications, we looked at the IBM Tivoli solution but it was just cost-prohibitive.
Which other solutions did I evaluate?
In terms of Stonebranch’s agent technology and compatibility with other vendors, it's been very good. We migrated off of the SAP platform or BOE; we had an old instance of Data Integrator. There were a lot of things that the other tools that we looked at when we did our PoC - tools such as Automic and Automate It - just simply wouldn't do. They had problems integrating with the old version of Data Integrator. We PoC'ed three platforms before we made the decision to go with Stonebranch. Automic was one, who was subsequently bought, about four months after we made our decision, by CA.
We looked at a Windows-based solution called Automate which wasn't anywhere close, at the time, to Stonebranch as far as schedulers go. Automate had a lot more UI-manipulation capabilities, which Stonebranch and Automic didn't really have - you don't usually go in and automate processes with the latter two. They're more straightforward workflow tools, rather than automating UI processes.
From our experience in the PoC, Stonebranch's web-based platform wasn't converted from a client-server platform like some of the other platforms we looked at. And being agent-based, it was easy to use. It didn't require a steep learning curve.
And, again, that integration to some of the technologies that we were using was important. We're now about 75 percent of the way to migrating all the Data Integrator services and some of the SSIS ETL jobs we have running out there to Informatica. Stonebranch integrated to all three of those ETL platforms right out-of-the-box through the REST API. We really didn't have to struggle with those integrations, so that was great.
What other advice do I have?
My advice applies to the implementation of any automation platform. Have a solid plan for what your approach is to automation and develop a common approach to workflows. That is what we did. We took a step back and looked at what our workflows actually do. We really only have four fundamental workflows so we built templates for those, and all but a couple of our workflows fit into those templates. So analyze your workflows, understand what they do, build templates to do that, and then go off and do the automation once you've got the templates built up.
I'm not getting any complaints from my automation team about the platform. The guys are really happy with it.
I really haven't spent a whole lot of time in the Stonebranch Marketplace.
We looked at Stonebranch's file-mover product for internal file movement in our workflows where we're moving files. We did a PoC with it and we decided to go with another product. We have three managed file-transfer platforms in our enterprise. The biggest one that we move most of our files with is IBM's Sterling File Gateway. One of the biggest challenges there, when you're moving data across the enterprise and doing workflows, is managed file-transfers platforms. They drop them in in the landing zone where processes will consume them. What we found was that those files were just sitting there waiting for somebody to manually do something with them. In the meantime, we've automated all that. We moved it from the landing zone to a retention zone where they sit in tier-two storage for 60 days. We then move them off to the data domain, which is tier-three storage, archives. The automation really helps slow down the eating up of tier-two storage platters at $40,000 a pop. Everybody's has this problem, but we've automated that process. We've even automated deleting those files off of tier-three storage and realized a savings with our Storage Spend.
We have an automation team of three employees. We have another six members of that team who are offshore contractors. We have four users and the analytics team which are actually doing workflows on their side as well. The way that works is that we have a dev, a test, and a production environment with Stonebranch. We have a Controller running in each of those environments. The analytics team builds its workflows up in dev. We do basic testing there and then push them to the test environment where the automation team makes them work. Then we push into production. We're mostly working with the analytics team, the dev environment, with them developing their triggers there.
For deployment and maintenance of the solution we need three people on the automation team.