Pushing the capacity limits
Pushing the capacity limits
March/April Edition, 2008: For many enterprises, the greatest data show on earth isn’t hidden away in some obscure marquee tent, but rather within the vast data stores where the lifeblood of the organisation flows. What are the limits for a massive organisation like Amazon.com?
When analyst firms like IDC and Gartner are putting you as the top dog in the data warehouse business, it’s easy to believe that Oracle hold the market leading position by a fair margin. This strong position has been helped along with a focus on clustering, and its ability to not be slowed down by the wait for technology.
“Enterprises run into the problem of what to do when they have the biggest and best server, if that maxes out then where do you turn?” says Peter Thomas, senior director of technology solutions at Oracle.
Clustering data is looked upon with hungry anticipation thanks to the ability to build flexibility, which is exactly how Amazon.com started to view the technology when they ran into data issues. Thomas explains that the online retailer ran into data warehousing issues with their “Also Bought” system where shoppers could see what other merchandise people who bought one item were also clicking onto their credit cards.
According to Thomas, the Seattle based company ran into a technology cap where they were running the data off the best available servers and were keen to find an alternative to waiting, “Amazon moved from huge Unix servers to small, clustered Intel servers running Linux,” he says.
“One of the problems Amazon ran into was Linux Volume Manager not being able to give the performance they needed so they switched to Auto Storage Manager which gave them double the Input/Output capacity.”
Another benefit that came from this process was increased auditing capabilities, which allow administrators to see who accessed what data, when it was and from what IP address. This is becoming of particular importance for organisations with a presence in the United States (US), Thomas explains that should you touch any part of the US shore as part of your business you need to be considering the Sarbanes Oxley style requirements for data leakage reporting.
Oracle has been showcasing the flash new features of their 11g database but not everyone has made the switch. While there is reluctance from some quarters to get on board with the new version, Thomas explains that the feedback has been positive from all those who have.