Today, many companies still have most of their transactional data in relational database management systems such as Oracle which support various business-critical applications, from order entry to financials. But in order to maintain processing performance, most companies limit the amount of data stored there, making it less useful for in-depth analysis.
One alternative, according to a recent DBTA webcast presented by Bill Brunt, product manager, SharePlex, at Dell, and Unisphere Research analyst Elliot King, is moving the data to Hadoop to allow data to be inexpensively stored and analyzed for new business insight.
A prevailing thought within the industry now is that companies can store their data in Hadoop and have all of their data in one convenient data lake. The advantages to this would be that all of the data is stored in one place. In theory, this would shorten the time it takes a company to analyze the data and make the data beneficial faster.
However, there are some disadvantages as well. Hadoop is a general tool that is made to work across all different data spectrums but is that better than more specific optimized tools. Also, with all of the data stored in one place data theft becomes easier for hackers. The challenge for Hadoop is how it should be integrated across data warehouses and other methods of storage. The three common ways are dual system, federated system, and single logical systems. King explained that companies that are considering moving data to Hadoop must make sure the business benefit outweighs the cost.
Dell an end-to-end solution for converting data from data warehouses to Hadoop, and has built reference hardware that is able to support multiple distributions of Hadoop. SharePlex is a lightweight and fast answer for getting data out of databases and making it available for other uses. “We see a lot of alignment within the industry for our strategy of data replication,” said Brunt, noting, SharePlex is similar to Hadoop in enabling low cost and high performance for the customer.
To access a replay of the webcast, “Moving Data from Oracle to Hadoop to Obtain New Business Insights,” go here.