News Stay informed about the latest enterprise technology news and product updates.

Amazon's Jeff Barr digs into AWS pricing, uptime on user group tour

At the first stop on his cross-country speaking tour, Amazon's Jeff Barr answered questions on AWS pricing, performance and availability.

Not many major corporations would be comfortable putting one of their more visible executives in front of an audience prepared to openly grill him without any public relations minders on standby.

So, either Amazon Web Services Chief Evangelist Jeff Barr makes his PR department a bit squeamish, or the cloud services giant is actively doubling down on customer outreach with Barr's road-trip tour of 14 AWS user groups around the United States, which kicked off in Cambridge, Mass., on Monday night. Although he came with a presentation, he frequently paused to field questions and feedback from the audience on AWS pricing, performance, availability and other topics during the two-hour event.

Barr's willingness to engage customers directly makes sense, given that cloud providers continue to struggle with getting their message out to customers. In a recent survey of 288 service providers, 37% indicated that educating customers about public and hybrid cloud was one of their main business challenges.

The need for education and awareness is not lost on Barr, whose presentation provided a fairly extensive walkthrough of the AWS portfolio -- highlighting well-known services like Elastic Compute Cloud and Simple Storage Service (S3), as well as some less-visible ones like AWS static Web hosting and Trusted Advisor, a management tool in beta that does a customized analysis for each user that identifies cost inefficiencies, performance bottlenecks and security holes.

He also highlighted Amazon's growth -- noting that the number of S3 objects has grown from 2.9 billion in 2007 to 2 trillion today -- as well as such high-profile AWS deployments as the AWS architecture used for President Obama's re-election campaign website.

"[People] came here because they wanted to learn," said Barr, who runs the eponymous Amazon Web Services Blog. "I love the fact that they're going to ask me questions I can't answer."

Barr, who will be driving 5,000 miles across the country for the next four weeks of his "AWS Road Trip," said the tour is an antidote to the lack of time and intimacy he has with customers on the conference circuit.

"Last month, I flew from Seattle to Beijing to speak for 15 minutes," he said. "I really, really love speaking at user groups. I find that user group attendees are always totally motivated guests … [and] I love the fact that we have time to have a really in-depth discussion."

While Barr faced a friendly audience, members of the Boston Amazon Web Service Meetup Group rose to that challenge, many calling on him to explain the ins and outs of several AWS pricing policies, including those for spot instances, reserved instances, Elastic Load Balancing (ELB) IP addresses, Elastic Block Store (EBS) and Amazon Glacier. Others asked about the efficiency of Amazon's application programming interfaces, the speed of EBS snapshots, and how customers could simulate failures.

AWS users raise concerns about availability

Availability has been a recurring concern at the user group's other meetings, said David Ghikas, cloud manager of Cloud Cowboys, a systems integrator in Andover, Mass., specializing in implementing and managing Microsoft products in an AWS environment. Due to the high concentration of customers in AWS' Virginia data center, its only facility on the East Coast, users are concerned about other regions becoming overloaded if a failure there caused a massive, collective failover, Ghikas said after the event. It's a concern that affects what he suggests to clients, he added.

"Because of the perception and the issues [surrounding that region], I recommend to customers that they have a failsafe copy [elsewhere] or to go to a different region," he said.

Only one AWS customer at the event raised a question about availability, however, asking Barr to clarify which AWS services are so closely tied that an outage in one would bring down the other.

"I would like to understand what AWS services depend on other AWS services, so I could think through different failure scenarios -- like how ELB, at one point, depended on EBS," the audience member said. "Separately, can I simulate types of failures that I might expect?"

"I don't believe we currently publicize the internal dependencies, but we're always working to get availability higher and higher," Barr responded, adding that he has asked the AWS team to develop a way for users to "simulate brokenness." Audience members chimed in that Netflix's Chaos Monkey, Chaos Gorilla and SimianArmy testing applications help meet that need.

Let us know what you think about the story; email Jessica Scarpati, site editor, and follow @jscarpati on Twitter.

Dig Deeper on Managed IaaS

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.