Software Engineering Institute | Carnegie Mellon University

SEI Insights


SEI Architecture Technology User Network (SATURN) News and Updates

SATURN 2013 Cloud Computing Session (notes)

Posted on by in

Notes by Frank M. Rischner’s Cloud Architecture
Joel Crabb, Best Buy, Inc.

Crabb works for BestBuy, the world largest multi-channel consumer.

In 2010, BestBuy built a team to experiment with cloud components. Smaller web properties have been stored in the cloud. Also, the test environments have been put in the cloud.

BestBuy uses caching for their traffic. The importance is flexibility, scalability, and reliability. Clouds fail, so 100% availability is important. Load balancers are used to balance the workload in the cloud. The product data is loaded from the data center into the cloud.

The front end is built by classic HTML, JavaScript, and CSS. It uses a templating framework. The front end had to be flexible. The data is transmitted from the back end to the front end using a JSON data contract.

Best Buy makes use of a NoSQL product datalog, which is in this case a key–value pair. Best Buy used distributed service frameworks.

Automated Provisioning of Cloud and Cloudlet Applications
Jeff Boleng and Grace Lewis, SEI
Vignesh Shenoy, Varun Tibrewal, and Manoj Subramaniam, Carnegie Mellon University

Jeff served 25 years in the Air Force, where he was involved in security concerns.

Motivation and goal:

Mobile computing elements that make use of architecture. Those mobile components should be automatically provisioned, secure, and assured. The approach was to demonstrate a full, end-to-end solution. It made use of a digital container for computing components, which is in OVF format.

Architectural overview:

The mobile device runs a surrogate client that connects to the surrogate, which is a cloudlet. The client sends the data to the cloud, and the cloud does the computing and sends it back to the device.

Surrogate and automated provisioning:

The component creator determines the optimal and minimum execution parameters while packaging.  There are multiple surrogates. Tests have been done with Java and Python containers doing face detection.

Use cases:


Included distributing search logic to distributed databases (moving the computation to the data), cyber foraging, and crowd analysis.

Applying Architectural Patterns for the Cloud: Lessons Learned During Pattern Mining and Application
Ralph Retter, Daimler TSS

Daimler is a German corporation producing mainly motor vehicles.

In 2010, the cloud-computing hype reached Germany, and every company wanted to get involved in cloud computing. So Retter and his team researched patterns. The goal was to answer some of the reoccurring questions about cloud computing: Which infrastructure to use? Does the application fit the cloud? Etc.

There are two approaches:

The bottom-up approach and the top-down approach. The real requirements do not always that mean we need the cloud. The top-down approach first looks at the business requirements and then at the application to the infrastructure.

Lessons learned:

First lesson learned: Modularize the systems to scale them up and out individually.

Availability is a big part of the patterns.

More lessons learned: “If you can’t give back resources – it ain’t cloud.”

The catalog of all the patterns can be found at


About the Author

Bill Pollak

Contact Bill Pollak
Visit the SEI Digital Library for other publications by Bill
View other blog posts by Bill Pollak



We welcome comments with a wide range of opinions and views. To keep the conversation focused on topic, we reserve the right to moderate comments.

Add a Comment


Type the characters you see in the picture above.