Prioritizing Principlas in "On Designing and Deploying Internet-Scale Services by Brad Porter. Brad highlights some of the most important ideas in implementing internet-scale services: 1) Keep it simple, 2) Expect failures, 3) Support version roll-back, 4) Maintain forward-and-backward compatibility, 5) Give enough information to diagnose. 3&4 are tough tough tough. Some comments by readers as well.
Waypoints From Big Data to Smart Data. Krishna Sankar thinks a transition from Big Data to Smart Data is happening...Smart Data = Big Data (data at scale, connected and probably with hidden variables) + context + embedded interactive (inference/reasoning) models.
Greg Linden with some good hits from his latest reading list. Topics include Google on social, Yahoo raising chickens, Google on VLDBs, the doom of both netbooks and tablets, YouTube becoming the TV of the future.
Christina Warren with How Groupon Uses the Cloud to Scale Its Business. I did not know that they used Salesforce and its Force.com to power their deals. Good description of the deal process too.
Microsoft with an interesting new graph database called Trinity. Trinity is a graph database and computation platform over distributed memory cloud. As a database, it provides features such as highly concurrent query processing, transaction, consistency control. As a computation platform, it provides synchronous and asynchronous batch-mode computations on large scale graphs. Trinity can be deployed on one machine or hundreds of machines.
Now for something completely different, TalkTechToMe will read HighScalability.com and other tech feeds to you over a phone.
George Beech with a StackExchange experience report on Performance tuning Intel NICs. By tuning their NICs peak traffic went much smoother. Have you tuned your NICs today?
Google releases snappy, a compression/decompression library. A point Google has made in the past is that cheap compression can double your bandwidth with very little cost.