Print this page
Published in News

Amazon mocks Oracle’s click bait story

by on25 October 2018


We must admit they have a point

Yesterday a story appeared on the wires based around a leaked insider report claiming that Amazon’s Prime Day was ruined by the outfits transfer from Oracle to Amazon’s software.

The story itself was credible, but alarm bells started to ring when we saw the original. The report appears to have existed but the question of how CNBC obtained it, whether they had seen the complete report and why the whole story was slanted in favour of Oracle’s software.

It is an odd intro to say Amazon is learning how hard it can be to move off of Oracle's database software when experts quoted in the same story told the opposite. CNBC’s version even had a nice quote from Larry Ellison which would have taken some time to organise.

Amazon has been in touch with us and said that the story was rubbish.

Amazon CTO Werner Vogel said that he had tried to help the reporter get it right, but clickbait won. In other words, the story is less interesting than it was pitched to CNBC which appears to have run it anyway (to be fair so did we).

“Our Fulfilment Centres have migrated 92 percent of DBs from Oracle to Aurora with better avail, fewer bugs and patches, less troubleshooting and less hardware cost”, Vegel said.

Vegel said that AWS and Aurora had absolutely nothing to do with the Amazon Retail Prime Day outage. The Prime Day website issue resulted from a problem in the Amazon Retail software stack.

The document that CNBC apparent got its paws on came from came from the Amazon Retail team and detailed a completely unrelated issue in a single fulfilment centre (out of more than 185 worldwide). This did lead to the slowing of processing in the fulfilment operations and a slight delay in the shipping of products from one facility.

“There was never an outage at the facility, and the issue only resulted in delaying shipping of about one per cent of packages for a brief period (unnoticeable to customers)”, Vegel said.

The application in question accidentally created an excessive number of savepoints, despite that team knowing that Aurora and Oracle handle savepoints differently.

This created a temporary situation where the database was slow, and the application experienced intermittent timeouts. The problem was quickly diagnosed and entirely resolved by merely removing the unnecessary save points that had been inadvertently left in the retail application. No changes were needed in Aurora.

According to Amazon, the move away from Oracle’s database system is working well, and it is pleased with the result. It would be nice though not to have your former vender slagging off your choices in the mainstream tech press when you don't decide to renew your licences.

 

 

Last modified on 25 October 2018
Rate this item
(0 votes)