Until recently, many organizations were able to conduct day-to-day operations without providing their employees with real-time visibility into data ensconced in various silos. However, there are a number of urgent business considerations that mandate more timely interaction with vital enterprise information, regardless of where it’s stored, including:

• Ever-expanding data silos: Organizations continue to buy and build new applications; mergers and acquisitions also add to the data portfolio.
• New regulatory requirements: Events of the past few years are mandating better visibility into the entire enterprise, in real-time.
• Globalization: Enterprises are now conducting business around the clock, around the world.

Even though these are all persuasive justifications for enhanced access to information, budgets have not kept pace with these demands.

Many IT organizations recognize that meaningful silo consolidation isn’t on the horizon. This means that they’ll continue maintaining multiple sources of information. Unfortunately, training and support budgets have been squeezed, which makes this task even more daunting. Happily, commodity hardware and networking capabilities have greatly expanded in recent years, with corresponding cost reductions. This opens the door to more innovative yet lower priced data distribution options.

Options for disseminating information

Since each passing day increases the likelihood that you’ll need to find a way to circulate your data, let’s look at some of the alternatives at your disposal.

• Rewriting applications to use the same data storage repository: One way to surmount the need to distribute data is to have all your applications simply use a single data storage repository from one vendor. However, given the costs, complexity, and amount of time necessary for this approach, it’s not very likely for most organizations.
• Distributing data using batch-oriented technology: Extract, Transform, and Load (ETL) technology has been transmitting information for many years. Unfortunately, in most cases, this approach doesn’t place timely data in users’ hands quickly enough. The result can be erroneous decision-making based on stale knowledge. In addition, this technique requires new software and skill sets.
• Employing real-time data integration: As we’ll see, this is the optimal approach to disseminate data in support of today’s fast-paced, globalized enterprise. There are two distinct strategies that you can use for this option:
  • Specialized, proprietary middleware: While this software does indeed move information across silos, it can be expensive to purchase and deploy, and requires protracted training to be effective.
  • Database-driven heterogeneous replication: As we discuss in the balance of the article, this approach is cost-effective, integrates well with your current infrastructure, and uses the skills that are already present in your organization.

Database-driven real-time heterogeneous replication

For most practitioners of this selection, data is kept in its native silo; copies are then distributed in real-time to other silos. Using a heterogeneous, log-based, bi-directional technique, data is created in one location and then distributed to others. Importantly, this is all handled without the need for user intervention. Administration takes place from a central location.

Since data distribution is bi-directional, this means that changes made anywhere will be reflected everywhere. In addition, remember that since this is a heterogeneous strategy, data is often kept in database engines from multiple vendors. For example, it’s common to see Sybase ASE data copied to an Oracle server, or DB2 information shared with a Microsoft SQL Server instance.

There are substantial benefits to opting for heterogeneous data replication. From a cost management perspective, this route protects your investments in existing applications and technologies, while imposing minimal impact on your infrastructure. For example, why replace a well-running eight-year-old packaged application simply because you need real-time access to its data from a brand new, internally developed solution? Heterogeneous data replication lets you copy, in real-time, pertinent information from your existing application into the corresponding destination in the new solution. Any changes made to the data elsewhere will instantaneously be copied back to the original silo. What’s especially compelling about this method is that it doesn’t require any modifications to the underlying software or infrastructure that’s supporting both the existing and new applications. Furthermore, most replication technologies that are well-integrated with the database platform don’t require IT administrators to learn entirely new sets of skills to support replication.

From the vantage of the business, heterogeneous data replication lets users continue to work with their suite of existing, well-understood applications. From the users’ perspective, it appears as if their applications have been extended with additional, real-time information taken from other silos. This broader set of information helps enhance decision-making and productivity, while improving operational efficiency. Finally, since most enterprises are very cost-conscious today, database-driven heterogeneous replication is a cost-effective solution with an impressive Return on Investment (ROI).

Central Vermont Public Service and heterogeneous replication

For a real-world example of heterogeneous data replication in action, consider the case of Central Vermont Public Service (CVPS). Organized in 1929 by the consolidation of eight electric companies, CVPS is the largest electric company in Vermont and serves close to 160,000 customers. Most of the firm’s database applications are home-grown and run on technology from a single database vendor. Their largest solution, Workforce Management, handles day-to-day service operations. According to Wendy Perry, lead database administrator at CVPS, “Through Workforce Management, we manage all of our service calls to make sure our customers properly receive electricity services.”

This Workforce Management database helps CVPS manage projects such as power line extensions, new designs, new construction, crew scheduling, material tracking, and personnel service-incident hours. As is the case with many enterprises, CVPS also relies on a database platform from another vendor. For CVPS, this involves an Intergraph application running on a database from another provider that helps the firm manage electrical facilities, lines, and poles.

The Workforce Management database performs as an online transaction database that constantly receives updates from approximately 100 users throughout the day while the database in support of Intergraph is used by a handful of users that run queries to generate reports. A key requirement for CVPS to operate its business properly is to ensure that the two databases share information efficiently. CVPS needs to keep the two databases separate but also needs them to bi-directionally exchange information updates.

At one point CVPS considered rewriting one of the applications so that IT would not have to support two major enterprise databases. But the time and cost to convert the Workforce Management application to the other database platform would have been even more exorbitant since the process required a complete rewrite rather than simple conversion. In addition, the Workforce Management database was also connected to other databases from the same vendor. These other databases were also vital pillars of CVPS’ business operations, which made the rewriting option even less palatable.

It was then that CVPS began looking at whether replication would meet their needs. “We realized replication was the best way to go” Perry reveals. "We simply keep the tables stored in the Intergraph application’s database and then proxy them to the Workforce Management database. The replication software includes all the middleware necessary for cross-database platform translation.

CVPS tested this approach and found it could properly replicate tables back and forth between the two systems. This interested the company because it would be a lot less expensive than a comprehensive middleware solution. Replication also offered a viable option because CVPS needs to replicate data in real-time to keep the Workforce Management application running and to ensure customer service calls are handled efficiently.

“By integrating our databases, we reduced our overall licensing costs,” Perry says. “A comprehensive middleware application would have been more costly and would not have performed as well when we tried to join tables across platforms. By using replication to transfer tables, we can keep both systems separate so that each database does not have to see tables through the proxy.”

The business users at CVPS are well-insulated from the intricacies of heterogeneous data replication. Instead, they’re reaping the benefits from instantaneous, cross-platform data access. For example, since the Intergraph application is responsible for managing physical assets such as light poles, transmission lines, and so forth, heterogeneous data replication grants Workforce Management users immediate access to this vital information. The result: faster, more productive service calls and far fewer power outages.

Business and technical realities will continue mandating timely and far-reaching information distribution. Given these certainties, it’s incumbent upon IT to deliver these results in an efficient, cost-effective manner. Whatever the chosen solution, it should protect investments in hardware and software, minimize new training requirements, and deliver immediate access to data. As we’ve just seen, heterogeneous data replication is an attractive choice that is worth investigating.

Robert Schneider is a Silicon Valley-based technology consultant writing for Sybase. He has written six books and numerous articles on advanced technical topics such as Service Oriented Architecture (SOA), Cloud Computing, and high-performance relational database design/optimization. He can be reached at This e-mail address is being protected from spambots. You need JavaScript enabled to view it .