Circuit Breaker solution in Netflix

The Netflix Tech Blog: Fault Tolerance in a High Volume, Distributed System
by Ben Christensen In an earlier post by Ben Schmaus, we shared the principles behind our circuit-breaker implementation. In that post, Ben discusses how the Netflix API interacts with dozens of systems in our service-oriented architecture, which makes the API inherently more vulnerable to any …

Facebook’s Go program combines two fairly new computational technologies: deep convolutional neural networks, which use vaguely brain-like algorithms to learn and recognize the importance of large-scale patterns on the game board, and Monte Carlo tree search, which is a form of “thinking ahead” that works out detailed tactical sequences of combat in specific areas. 

Why Go Still Foils the Computers
Facebook and Google are working to enable computers to play the game of Go, an especially complex game that dates back over 2,500 years

2014 – Big Data is Out, Machine Learning is in.

"Big Data" is no where to be seen the hype cycle 2015 mean that the most talked about big data related technologies are now into practice and no more a hype.

"Machine Learning" made its first appearance on the chart this year, but already past the peak of inflated expectations and now takes the place of Big Data.

The hype cycle gives us an idea of which of these technologies actually survive the market hype and have a potential to become a part of our daily life. The emerging technologies is based on the assessment of maturity, duration for adoption, business benefit and future direction of more than 2,000 technologies, grouped into 112 areas.

Five regions of Gartner's Hype cycle:

• Innovation Trigger (potential technology breakthrough kicks off)
• Peak of Inflated Expectations (Success stories through early publicity)
• Trough of Disillusionment (waning interest)
• Slope of Enlightenment (2nd & 3rd generation products appear)
• Plateau of Productivity (Mainstream adoption starts)


In album 1/1/16