In March 2019, HostBridge CEO Russ Teubner demonstrated the HostBridge Transaction Analytics Connector (HTAC). The demonstration showed how hybrid IT organizations can gain end-to-end visibility of hybrid application performance when those hybrid applications include CICS.
HTAC is software that correlates non-mainframe service requests to the actual work done inside CICS to process those requests. By using HTAC, organizations that rely on IT Ops platforms like Splunk or Elastic can understand the performance of a transaction that begins in the cloud or on a mobile device, goes up to the mainframe to invoke CICS transactions, and then returns a result to the end user. With the complete, end-to-end visibility that HTAC enables, IT organizations can better manage the performance of their hybrid IT applications. This is exactly what one auto-insurer did with HTAC, discovering a way to eliminate 28 million unnecessary instructions that executed each time its most common business process was invoked.
Gaining End-to-End Visibility is Difficult
During the webinar, attendees were asked to rate the difficulty of knowing how web, cloud, or mobile activity is driving mainframe CICS transaction workload. The results of this poll are shared here:
No one reported that it was easy to know how web, cloud, or mobile applications drive mainframe CICS workload.
As the graph shows, none of the webinar participants reported that it was easy to know how web, cloud, or mobile applications drive CICS workload. In fact, one attendee, in response to the poll, volunteered the comment that the response option of “nearly impossible” was missing!
The recorded webinar is now available for viewing on the HostBridge YouTube channel. View it here:
The presentation included a Q & A session, which is summarized here:
Q: What SIEM platforms does HTAC support?
A: Even though the demonstration used Splunk, HTAC is not a Splunk-specific solution. When we were developing HTAC, we surveyed a handful of customers, and they were all using Splunk, making support for it a reasonable starting point. But we have other customers who are interested in platforms like Elastic, and even others beyond that. What we have found in our customer base is that the interest is in Splunk first, and Elastic second, and HTAC supports them both.
What we are doing with HTAC is ensuring that this request correlation information is added to the data that CICS ultimately produces for a transaction. Therefore, any mechanism that a customer can conceive for getting data into an analytics platform – perhaps even something they developed in-house – HTAC can support. HTAC emits a JSON document, which will allow HTAC to support any analytics platform that can process a JSON input.
Q: CICS is the tail end of a much larger message flow, starting potentially at a home internet or mobile user. Do you have a way to tie correlation data that starts in a distributed system so we can see the full picture (i.e., ITCAM, etc.)?
A: This scenario is exactly what HTAC does. Requests that originate with a home user, either through the internet or a mobile connection, will go through some intermediate server. Then, that intermediate server may route the request through another server, and eventually the request gets to CICS on the mainframe – the “tail end” of the message flow. As long as that point of contact interacting with CICS can put something in the request flow – in a channel, container, COMMAREA, request header, an MQ message, or somewhere – then HTAC will extract that correlation data, enrich the Transaction Tracking origin data and it will come through.
There are certain scenarios where it is just impossible for the client device or application to get a common set of correlation data all the way through to CICS on the mainframe. It’s just too hard or logistically impossible. In these situations, because of the power of Splunk or Elastic, all a user has to do is take the correlation data from each separate leg of the message flow, merge them together, and get a composite view. To summarize, this question is exactly what we built HTAC to do. It is a “targeted” strike that makes sure that the “last mile” – the leg of the message flow that goes into the mainframe – is correctly annotated with request data so that the user can piece together a complete picture of the entire message flow.
Q: Is it necessary to modify application code in order to use HTAC and the Splunk solution?
A: Absolutely not! There is no need to modify applications to use HTAC. When HTAC does its extraction work, whether it is coming in across HTTP, CTG, or Sockets, there is no need to modify the application.
There are certain situations we can envision with customers who have correlation data inside of their applications that they wish to add through HTAC to the origin data. We’ve already had one customer that wants to do this. What we have done to accommodate this scenario is provide a programmatic call or exit that customer programs can use to enrich the origin data directly. But it is important for customers to know that, out-of-the-box, there are no applications changes required to take advantage of HTAC.
Q: Does HTAC have any other HostBridge software as a prerequisite?
A: No, HTAC does not require any other HostBridge software.
Q: How does your product compare to other end-to-end response time analytic engine products for tracing the entire flow of a transaction?
A: There are products that focus on end-to-end response time. The problem, however, is that their granularity on the mainframe seems to breakdown quickly. In other words, it is easy for these products to see an HTTP request sent into the mainframe and the response, and then add that response time to the overall response time grid and report it to the user. What this does, though, is essentially keep the mainframe as a “black box” on that response time grid. With HTAC, we are able to go far deeper.
Customers that have an interest in HTAC are not content with this view of the mainframe as a “black box” that gets a request and then responds in a given number of milliseconds. They want to know what’s really happening on the mainframe, tracking end-to-end the transactional elements of the workload on the host that were executed, and what performance impact they had. The difference is that with HTAC, we’re providing far more resolution about what is happening on the mainframe, as well as showing performance of the entire transaction flow from point of origin. Furthermore, you get to see this on a platform like Splunk or Elastic.
HTAC is available for testing at no cost to customers. To request a copy of HTAC, or to learn more about it, contact us using the form below.
Complete this form to start a dialogue with HostBridge about analyzing your CICS resource consumption: