Operational responsiveness is the ability of enterprises to closely track all business functions, flows, and be able to sense, and quickly respond to exigencies and unexpected business exceptions, says Ramesh Loganathan, VP (Products) and Centre Head of Hyderabad-based Progress Software Development P Ltd ( >http://bit.ly/F4TProgress ).
In enterprises, multiple applications co-exist to deliver a business flow or process, and so this responsiveness involves both tracking transactional processes in the organisation as well as effecting structural changes as they get warranted with changing business and operating climate for the enterprise, in real time, he adds, during a recent interaction with Business Line . “This is a shift from present paradigms where business managers analyse what has happened instead of what is about to happen; wherein the time taken to identify a change or trend results in missed opportunities.”
Taking the discussion of responsive process management (RPM) under the hood, Ramesh explains that to drive operational responsiveness, organisations must: Monitor all relevant business events as they occur, wherever they occur, internally or externally; analyse these events in real time, using prescribed business criteria to identify opportunities, inefficiencies, and risks; visualise key business indicators on real time dashboards; respond to capitalise on opportunities, increase efficiency, and reduce risk by alerting decision-makers to important business events, performing automated actions and initiating response processes.
Excerpts from the interview:
Can you give an example to highlight the need for operational responsiveness?
Consider this scenario, in logistics. Operational responsiveness would mean ensuring that every shipment reaches the destination on time.
In this solution space, the shipping system has the details of the shipment, the fleet tracking system has the details on the location of the trucks, a hand-held terminal may have the details of the actual field delivery of the shipment, and probably a weather tracking system provides inputs to both the controllers and the trucks on weather conditions.
While all these solutions function together to serve the common business objective of ensuring the shipment is delivered in the assured time, there may not be any automatic co-ordination or orchestration. Often, even as each of the four solutions may be automated, the integrated analysis and actioning are performed by human operators.
It would, therefore, be useful to business if the flow is automated, so that one realises the ability to sense and respond immediately to the events that impact businesses, using business-focused, event-centric infrastructure solutions that deliver operational responsiveness. That way, business can enable the right things to happen, at the right time, every time.
Are there different ways in which responsiveness is made possible?
Non-intrusively, getting access to business service invocations, business events and business flows/ processes involves, in some form, being able to intercept the same as they occur in the enterprise.
Typically in servicing any business process such as a new order process, the request may come to a web page or a web service (probably implemented as a servlet, running on a web server) and flows through several applications. All these may either happen in one coordinated business flow, or could happen through a combination of automated flows or end users actually working directly on the UI of different applications.
Automated process in its best form may be in a business process management or process orchestration engine. In the not-so-automated forms, this may involve accessing web services or accessing different applications through adapters. And in the least-automated form, the task may involve end users directly accessing different applications and performing the needed processing in each.
Under the hood, RPM would include capabilities for transaction insights, event capture and non-intrusive process discovery. The complex event processing (CEP) engine takes care of correlation; the rules engine assists decision-making; BPM (business process management) works at executing downstream business processes; and connectivity capabilities enable data and services integration.
As for interface with users, the ‘business control tower’ can be a configurable framework with visibility into key performance indicators (KPIs) and the ability to raise alerts, and act on them, in real time. Analytics process information and provide continuous real time insight to facilitate better decision-making. And enterprise integration enables interoperability of heterogeneous systems within a distributed environment.
What are the challenges in building a solution?
In realising RPM, some key technical challenges that will need to be addressed are:
* Non-intrusiveness:
-- Discovery of the systems, applications and process flows in the solution space.
-- Non-intrusively intercepting applications and inter-application or system interactions to extract business events and business data as the business flows and transactions occur.
* Automatic correlation of flows:
-- Linking up the business events or business function/ service invocations across a particular business flow. (This enables tracing a particular business processing, such as the servicing of ‘one order’ or ‘one shipping’ request.)
-- Correlating more complex business events across loosely coupled domains of applications, such as weather information off the web, GPS data of a fleet being tracked, warehousing, and shipping solutions that the enterprise is using.
* Dashboards for visibility:
-- Visibility to the business process insights gained by the solution.
-- Simple mechanism to be able to create the dashboards.
-- Alerting mechanism to trigger and process alerts.
* Actions:
-- Downstream remedial processing, maybe a business process or orchestration engine.
On achieving visibility non-intrusively.
The most difficult part of the solution is the business event and transaction/ process capture, especially if this is to be realised without requiring any (or much) changes to existing applications in the enterprise.
Towards a non-intrusive approach to capturing business events, the visibility model understands the list of applications, the business functions/services in each application accessed from outside (either users/other applications), the sequence of such invocations across applications that occur in the context of one business operation, the business information that flows across these invocations (say, an order or a shipment), the model for the business process under which the above are being executed, and the mechanism to detect the expected business parameters as each service or process step executes.
In distributed enterprise IT infrastructure, the interconnections between the various applications involves invocations across the applications that span multiple machines; each time as cross-app invocation needed, the source application composes a request and transmits the same across the network to the target application and the same gets processed at the target machine.
To get the visibility at each application, there needs to be a framework that allows understanding all requests coming in from outside that application; and, likewise, understanding all requests going out from that application to other downstream applications. In essence, a mechanism to understand all incoming requests into the application, the responses back, and any cascading downstream requests from this application that may be sent as part of servicing the said application request.
Getting insight into the individual steps/biz operations is only the beginning. Once that is done, the solution will need to correlate the various parts of a given business flow. Essentially, this calls for analysing various such data points and identifying those that belong to a single business flow. In environments that are not fully automated using SOA orchestration or BPM processes, this correlation is easier said than done.
It is important that the solutions do not in any way hamper or impede the execution and performance of the applications. Technologies have evolved to facilitate RPM without any loss of integrity to the request, without adding any noticeable latency, and without consuming any additional CPU or memory.
**
>InterviewsInsights.blogspot.com
