 | Issue: | Asia-Pacific II 2013 |
| Article no.: | 5 | |
| Topic: | Event-enabled enterprise – better and faster | |
| Author: | Kevin Pool | |
| Title: | CTO | |
| Organisation: | TIBCO Asia | |
| PDF size: | 229KB |
About author
Kevin Pool is the CTO at TIBCO Asia where he is responsible for helping client companies envision and execute their corporate transformation objectives. Prior to joining TIBCO, Mr. Pool lead the development and integration organizations at Great-West Health Insurance and at 360networks. He also built and led the professional integration services team which implemented order management and provisioning systems for Sprint, Level 3, Williams and WorldCom.
Kevin Pool holds a BS, Robotics, Western Washington University, and a MS, Space System Management from Webster University, Colorado.
Article abstract
Today’s inexpensive memory lets us go past yesterday’s memory bound paradigms and process events as well as data. An event-enabled enterprise’s information systems and processes can detect significant business-related events and immediately process them. Correlating event information with other events, and applying configurable business rules, enables you to develop new insights into a vast variety of business issues in real-time. Instant in-memory access to related information allows you to decide and respond in real-time, which provides tremendous improvement for your business.
Full Article
Your customer is giving you information. Are you listening? Are you responding? Your customer is using their mobile device to browse their usage and their minutes remaining. Why? Are they thinking about churning? Your customer is online and they are looking at your data plans. Why? Do they need more or less bandwidth? If the customer is thinking about a purchase or even churning, right now (while they are interacting with you) would be the best time to give them an offer – before they take the time to start shopping the competition.
Better yet, how about giving them an offer that incorporates everything you (should) know about that customer. An offer that is based on understanding what products they have, what they don’t have and even what products they had before and have since changed – an offer that considers the customer’s usage history and trends. How about incorporating location information? Where are they now? Where do they most use your products? The offer needs to take into account the propensity for the customer to accept this particular offer. And if you can give this offer to the customer right now, the likelihood of acceptance increases dramatically.
Your systems are also telling you things; are you listening and responding? Your remote network element has a sensor indication that is trending upwards towards an unsafe reading. Your field technician is travelling towards a scheduled customer call, and the automated tracking device in the truck locates the truck near the remote network element. The customer just called and cancelled the appointment. Can you correlate this information in real time and divert the technician to check up on the potential problem with the network element?
Business events are more than transactions; they are anything which occurs which could be significant to your business. We’re all familiar with transactions, but have you thought about the other events that are occurring continuously around your business? Most interactions with a customer (should) generate many business significant events. Yet we routinely only record the transaction and neglect to capture and incorporate all of the business events. Our systems can also be a rich source of event information, but we have to capture it first. Then we need to correlate the events to other events and incorporate other related information into decisions. Finally, we need to respond, and respond right now.
We are suffering from the hangover of how systems operated 20 and 30 years ago. Back then, processing was slow and data storage was expensive. Systems and entire architectures were built around the fundamental constraint that memory was precious and operations and IT staff had to allocate when and what made it into memory at any given time. We then forced business processes around the constraints of the systems that we gave to our users. Everything was structured around a ‘file cabinet’ pattern, where the database was our file cabinet and space in the cabinets was limited. It took time to put things in the cabinet and it took even longer to get things out of the cabinet. Then our systems and business process were all based around slowly filling up the cabinet with information and periodically digging around in the cabinet to finds things. How many of your systems and business processes are based upon 20-year-old thinking?
Over the past few months, I’ve had the opportunity to meet with five different mobile service providers in Asia, and three in India, Indonesia and Taiwan. The situation was pretty much the same for each of these operators. “The billing system is overloaded and it can take up to 30 seconds to look up information. The system is fragile, and we cannot afford to put any more load on it.” In a couple situations, the discussion was also about the customer relationship management (CRM) system. They all acknowledged the benefits of being able to respond in real time to their customers and for operational situations, but they just did not think their existing legacy systems could support it.
In modern data centres and computing architectures, there’s plenty of addressable memory and servers full of non-volatile flash memory available for applications. In-memory data look-up can be 100 times faster than a database. The price of computer memory has plummeted to less than five per cent of the cost a decade ago; some types now cost less than US$1,00 per gigabyte. Additional technologies are now available that lets us tie the memory of multiple computers together into a multi-node data grid, which can contain enormous data sets. With these capabilities, what’s happening fundamentally is a shift from the reason databases even existed. Don’t get me wrong, we will still needed databases, but the purpose of databases is shifting towards archival use only and the operational data that you run your business on is all being put in memory.
In telecommunications, we’re familiar with putting look-up tables and reference information into the memory of network elements to accelerate processing speeds and reduce to latencies required for providing telecommunications services. Even so, have you considered putting all of your customer, inventory and operations data into memory or about expanding this data to include whatever significant business events as they occur? This is now possible; not only putting all the data into memory, it’s also possible to have systems that monitor, correlate, react and respond to all of this information in real time. Having the data in memory, with systems integrated with this rich information, enables a more brain-like process with a matrix of information the computer can harness.
Powerful systems are now available which have the capability to process each of your business events as they occur. Event-correlation and event-processing systems use a distributed grid of servers, which interact with each other to process the large in-memory data set. Event servers immediately notify these applications when a business event occurs. Powerful rule-based systems allow you to configure, monitor and change to the system’s behaviour without requiring new builds or deployments from your IT department. Real-time dashboards and real-time analytics let you know what is happening with your business right now.
One major retailer in the U.S. has embarked on the journey of becoming an event-enabled enterprise. The sizes for the first systems indicate the magnitude of what is now possible. Information on 50 million customers and one billion data items are all contained in a 500 Gigabyte distributed in-memory data grid. Each second 2000 events are incorporated into the processing. Using this information, they are able to respond in real time to customers with offers specifically selected for that customer. Offer uptake has increased from 0.5 per cent to 14 per cent. That’s a 2,700 per cent increase! And these are just their first steps; there are lots of other projects in the works.
Of the five Asia mobile operators that I visited in the last few months, three of them (one in each of India, Indonesia and Taiwan) have started on the journey towards becoming event-enabled enterprises. They are restructuring their systems to generate, capture and respond to events. They are moving operational data out of databases and into memory. They are using event servers and rule-based systems to correlate and respond to business events. While each are approaching things somewhat differently depending upon their specific objectives, they all recognize the constraints of their legacy systems and are taking bold steps toward becoming event-enabled enterprises.
Just think what you could do if you had an event-enabled enterprise right now!
