AWS Architect Gives A Class On Cloud Building 101

The first week with the Amazon cloud is marked by a high-speed setup process that eventually gives way to incremental future adjustments as the cloud becomes more deeply embedded in services and procedures.

Amazon partners and customers attending a breakout session at the AWS Summit in San Francisco learned about how setup processes that took weeks or months in the old paradigm could literally be accomplished in five days. AWS solutions architect Don Southard walked attendees through a model of where to begin, and how to build upon the various available offerings and refinements.

Southard demonstrated that day one is primarily a matter of getting onto the system, setting up Amazon Machine Images, which essentially are used as templates for virtual servers, and engaging in a variety of configuration processes that relate to user access, machine images, instance sizes and security measures.

[Related: AWS Summit: New Cloud Avenues Open Every Day ]

"When you do this portion of the setup, you'll receive a key pair, which is very important to keep track of, especially when you're first starting up," Southard told the gathering. "Amazon does not keep copies of those."

Southard then walked through a lengthy checklist of similar activities that round out the remainder of the first week, including the setup of a tiered architecture, tool selection, failure planning, load balancing, and setting up the desirable levels of automation.

At the conclusion of the first week, following all the setup and testing, the focus turns toward applications and services that will deliver the ultimate business value of the cloud.

One solution provider attending the conference noted that as customers become more accustomed to cloud-based services, the rate of uptake will continue to accelerate.

"I think Infrastructure-as-a-Service will see an even faster adoption curve than we've seen for the rest of the cloud," said Russ Young, executive vice president of LTech, a Bridgewater, N.J.-based solution provider. "Customers are becoming far more accepting of moving toward utility-based compute models. Typically, they start with Software-as-a-Service, and then extend to Platform-as-a-Service. As they have success with those two models, it becomes a lot easier to extend it to their overall infrastructure."

Young added that he has trained his sales team to assess customer willingness to move to the cloud, but not try to force them to do so. "If they're not ready right away, they will likely be ready in the future," he said. "You might be talking to different people when that happens, but most companies will likely move in that direction."

NEXT: Search For Bottlenecks

One of the key elements that solution providers need to address involves continuous access to the applications.

"You've always got to be watching for the bottlenecks," said Rajakumar Sampathkumar, engagement manager for CSS, a San Jose, Calif.-based integrator. "You need to do a lot of benchmarking, and once bottlenecks are discovered, you can determine how to scale, or work around it."

Meanwhile, Bob Koche, vice president of SeeVogh, a solution provider focused on cloud-enabled videoconferencing, is highly familiar with latency issues.

"You've got to watch them closely and know how to make the necessary adjustments," he said. "Customers are happy to go with the cloud, but aside from latency, they're also very concerned about security. Whenever there's an attack on the Internet, we all suffer."

Many industry experts acknowledge that security will continue to be a moving target for information technology of all types, including the cloud. Amazon customers take responsibility for security above the hypervisor, while the company assumes responsibility for the lower portion of the stack.

PUBLISHED MAY 1, 2013