In-Depth
IBM Acquires Key Mainframe, System i Integration Technology
IBM’s acquisition of DataMirror’s technology will likely find its way into the company’s DB2 database
IBM Corp.’s acquisition last week of DataMirror Corp. means a lot more than just the exit of another independent data integration player.
DataMirror has a huge mainframe and System i (nee AS/400) presence. Its change data capture (CDC) technology for System z—i.e., DataMirror Transformation Server for z/OS—is one of the preeminent packages for that platform. It’s based on DataMirror’s log-based CDC technology, which can capture changed data from DB2’s database recovery logs. This helps circumvent the overhead and risks associated with trigger- and table-scan-based approaches, DataMirror officials say.
One upshot of Big Blue’s acquisition of DataMirror is that this technology will likely find its way—either as part of the standard feature set or, more likely, as a premium add-on—into IBM’s DB2 database. This should be a boon to Big Iron DB2.
DataMirror likes to position Transformation Server as a real-time (or "right-time") technology: Deltas are captured as they are written to the database logs and replicated to distributed RDBMSes or middleware message queues. In this respect, its capabilities mesh nicely with IBM’s acquisition strategy.
"Organizations need the ability to capture and use information in real time to help make better business decisions, better serve their customers, and increase operational efficiencies," said Ambuj Goyal, general manager of IBM Information Management. "The combination of DataMirror technology and IBM information management software will help customers bring real-time data analysis closer to actual business processes, allowing them to be more competitive and to generate more value from their information."
DataMirror officials claim that Transformation Server uses significantly less CPU time than competing solutions, largely because of its "single-scrape" capability, a DB2 log-parsing facility which ensures that only one read of the database logs takes place—regardless of how many integration processes are running concurrently. The result, DataMirror claims, is reduced demand on DB2 and the DB2 logs. This capability outstrips the barebones replication which IBM (and other RDBMS vendors) build into their databases, experts say.
"The replication tool market is odd, because there’s a big gulf between low-end and high-end replication," says Philip Russom, a senior manager with TDWI research arm. "For instance, all database management systems have replication built in, but it’s pretty basic stuff because it only works with that one database brand, and it’s usually for one-way data movement that copies data without transformation, and only between a pair of databases."
DataMirror offers sophisticated data replication capabilities, according to Russom: "DataMirror supports those rudimentary capabilities, but also advanced ones, such as native support for all the leading database brands plus legacy platforms, two-way data synchronization, data transformation, and configurations of many database instances."
A Good Fit
In this respect, Russom says, DataMirror’s CDC technology will be a good fit for IBM’s real-time information management push. "Replication, by nature, works well in real time," Russom notes. "After all, replication’s origins are in disaster recovery and database high availability, where you copy every single database transaction in real time as it happens to a mirror—or exact copy of the database—so that the mirror can take over if the main database server crashes."
Russom is not alone in this assessment. Charles King, a principal with consultancy Pund-IT, also applauds Big Blue’s move.
"The company’s purchase offers IBM distinct and ongoing benefits….DataMirror was considered a leader—along with Golden Gate Software, Lakeview Technology, and Vision Solutions—in its area of excellence, but Lakeview purchased Vision early last month," he comments. "The IBM deal should provide DataMirror the means to better compete against rivals in a rapidly consolidating market and also provide IBM stable access to leading technologies." There’s also the heterogeneous replication tip—i.e., the same differentiator that Russom himself touched upon. "[DataMirror’s] products support heterogeneous databases—including IBM’s DB2, Oracle, and Microsoft’s SQL Server—and operating system environments, including UNIX, Windows, and Linux," he points out. What’s more, DataMirror supports replication between and among Teradata, Sybase, and other data sources. "These features and functionalities will provide IBM ways to leverage DataMirror technologies to benefit its own multiple hardware and software platforms, and also to gain from sales of rival database solutions." Another selling point is DataMirror’s sizeable base of 2,200 customers—a roster that includes luminaries such as FedEx Ground, Union Pacific Railroad, and Tiffany and Co. Finally, King says, DataMirror just seems like a good fit for IBM’s overall data management—and Information Server—strategy.
"Of equal importance, DataMirror’s offerings appear to be an ideal fit for IBM’s Information Server product, which provides the basis of its greater [information on demand] strategy," he concludes. "By integrating DataMirror technologies, IBM will provide current and future customers the means to better discover, share, and leverage valuable information assets."