Q&A: Business Transaction Monitoring and Proactive IT Management
One way to keep problems from compromising your enterprise is to have a good monitoring system in place. We discuss the drivers and benefits of business transaction management, how it fits in your environment, and how your colleagues are using it.
IT's job is to keep an enterprises data systems flowing smoothly, and nothing helps more than to nip problems in the bud before they impact operations and, in turn, your customers. To learn how and why business transaction management (BTM) should be part of your monitoring toolset, we turned to Charles R. Rich, vice president of marketing and product management at Nastel Technologies (www.nastel.com), a BTM software solutions provider.
Enterprise Strategies: What does business transaction management (BTM) mean and what benefits does it offer? How costly can transaction errors become?
Charles R. Rich: Let me take your last question first. Transaction errors can be very costly. An order in a B2B firm could be worth millions, and it can run slowly or even be lost. Eventually, the customer organization may find the order and resolve the problem. However, both a major disruption to their business processes and inefficient usage of their staff would have occurred. In a highly regulated industry such as financial services, this becomes a major risk factor. Firms in financial services are very concerned about managing risk and providing effective IT governance. Business transaction management is designed to help address this by providing visibility, problem prediction, and improved transaction performance.
BTM is a dimension within application performance management (APM). It specializes in profiling the performance of transactions. There are two types of transactions: IT transactions and business transactions. BTM discovers and monitors the IT transactions that comprise a business transaction. For example, a business transaction might be something like "update order." However, in order for this business transaction to execute, many Java, .NET, and database calls must execute.
These activities are the IT transactions. The IT transactions occur in a specific order based on their dependencies as designed by the builders of the applications. BTM discovers these IT transactions and their dependencies, then tracks their performance as together they complete (or not) their parent business transaction. Performance of transactions is compared to what the business requires to remain competitive as specified in a service-level agreement.
Recently, the definition of BTM has been expanded further into something referred to as business transaction performance (BTP), which uses a complex event processing (CEP) engine to correlate operational metrics for servers and networks with transactions. Further, BTP does a deep-dive analysis of middleware messaging both in terms of transactions and payload contents.
All this expands the value of BTM -- "what happened" (transactional monitoring) and links that with "why it happened" (operational monitoring). The additional value of "why it happened" is critical, as this is actionable information that can be used to improve the situation and reduce time to resolution, and it enables IT to step up and rapidly deliver what the business needs to be successful.
How does BTM fit in with being proactive and preventing application performance problems and disruptions to business processes?
One of the best ways to reduce both risk and cost is to prevent problems rather than wait for them to hit your service desk before taking action. There is a saying, "Do you want to fix problems when they show up or when they blow up?" Clearly, we want to catch the early symptoms of these problems before a ticket is opened up at the service desk and resolve them, thus, avoiding all customer and business process impact.
Using a CEP engine, BTM can scan for patterns that foretell that we are heading for a "business abnormal" situation. IT can then take action before there is damage, thus, they are able to squeeze the stealth waste out of business processes that are often overlooked or deemed non-addressable.
For example, take U.S. banks. When complying with the requirement for reconciliation with the Federal Reserve by the end of each day, a bank's BTM system might, for example, detect 20 transactions in flight with perhaps an average duration of 10 minutes, each at around 4 PM. The BTM system with its CEP engine could determine that there is no way these transactions will clear in time. Strikingly, this happens at the very same time everything is "green" on the bank's monitoring dashboard and user impacts have not occurred.
You might believe that "all is well," but you would be wrong. If no action is taken, this bank will be in big trouble. Fortunately, the BTM system can alert its users that a problem will happen if they don't take immediate action to prevent it. Having the system in place allows bank personnel to deal with the problem directly or let an automated response to the problem occur.
Do you need to replace a customer's existing systems management tools that perform server monitoring in order to realize the benefits of proactive business transaction management?
A BTM system can be deployed using an "embrace" rather than a "replace" strategy. Think of it as raising the IQ of your existing investments. The BTM system can consume the alerts that the customer's existing application and system monitoring tools are harvesting and enrich that with transactional performance data. The new enriched information can be fed back to the customer's existing consoles and dashboards providing a seamless improvement process without having to throw away last year's investment.
What are the business benefits of tracking application transactions running in your enterprise or between yours and your business partners?
Very often, a B2B firm's transactions do not stay within the borders of its own enterprise. Whether the business is using SOA, enterprise service bus, or even the cloud, transactions will flow from a trading firm to the clearing house or from a manufacturer to its distributors. The forensic work of figuring out what happened and how to make it better can't stop at the company line. These businesses need the visibility to see their transactions run end-to-end across company boundaries to their trading partners and back. The benefits of this level of visibility are a reduction in risk, containment of cost, and, most important, competitive differentiation in terms of service as application performance is improved.
What are the root causes of transaction problems? Is running out of resources typically the root cause?
It is a common assumption that the root of most transaction performance degradations or failures ultimately turns out to be a shortfall of some key resource (e.g., network bandwidth, memory, disk, or CPU). However, practitioners of BTM have learned that quite often the root cause can be what is sometimes referred to as a "normal accident." This means the application and its transactions did what it was told to do (via its code); however, that isn't any longer what the business needs it to do. Sometimes this is due to a data-cleansing issue where a firm no longer settled a trade with company x; however, the data for the application was never changed in production. Having the visibility to determine that this is the problem when the phone rings and an important customer says "Where is my order?" is priceless.
Is BTM only for distributed computers systems and their applications or does it apply to both distributed and mainframe applications? Do customers with multi-tier applications need the benefit of transaction management and can they receive it?
Most Global 2000 firms have multi-tier applications running across Windows, Unix/Linux, and the mainframe. Similarly, their business transactions have components that execute on all of these platforms. In order to provide end-to-end visibility, it is essential that a user be able to trace transactions as they traverse each of these systems and not suddenly "lose the trail" when we hit the mainframe.
How are customers using proactive transaction monitoring and what benefits are they receiving?
Customers often start off using proactive transaction monitoring to reduce the backlog at their service desk and then move on from there. For the financial industry, BTM coupled with a CEP engine can keep transactional data-handling functions such as settlement, reconciliation, and compliance moving in real-time -- this is often used in the financial services "middle office." BTM helps financial organizations gain the appropriate visibility to enable rapid automated searches for patterns across large volumes of data to determine if they are at "business normal" or veering towards "business abnormal."
One large business we know has saved millions by reducing their severity 1 incidents. Severity 1 incidents are by definition those that stop the business in its tracks. By dealing with the symptoms before they become problems, these issues never blossom into Severity 1 incidents.
What types of organizations are typical users of business transaction management systems?
Usually Global 2000 firms with high messaging and transactional loads are good candidates for business transaction management. This is prevalent in the financial services, health-care, and insurance industries. It is also appropriate in high-volume retail.
Is there an appreciable ROI from this technology?
Yes. ROI from BTM is calculated by its reduction in both capital and operational expenditures. Many businesses try to solve their application performance problems by throwing computer hardware at them until they disappear … and are consistently surprised when this strategy doesn't work. In today's world of service oriented architecture (SOA) and cloud computing, "fixing" a server can have little overall effect on one of these "new architecture" applications. Plus, from a techno-economic standpoint, fixing slow applications by over-provisioning hardware simply doesn't solve the problem.
Between the economy and the abundance of technology available, businesses have made it a point to place heavier scrutiny on IT budgets, and needless spending is never a valid option, especially when the results are less than stellar. Avoiding this approach reduces capital expenditures.
Operational expenditures are reduced by improved staff allocation, decreased downtime, and a reduction in business process disruption.
What does Nastel Technologies offer to provide proactive business transaction management to customers?
Nastel AutoPilot is an application performance management software solution that reduces the cost of ensuring peak-performance of business applications. It provides businesses with 360-degree situational awareness of operational and transactional data across applications, transactions and messaging middleware. This ensures optimum service levels while also eliminating the hidden costs in business processes. At Nastel, we call these hidden costs "stealth waste." AutoPilot, with its correlation of transactional and operational data in a unified business context, enables management teams to identify and fix problems -- even predict and prevent them -- before they compromise SLA or other performance standards."