Today, you probably have a smart TV and it could well be the only other smart device besides your mobile and tablets. Tomorrow, you could be having a smart fridge, smart washer, smart vacuum cleaner, smart cupboards, smart beds, smart pillows, a connected grill station, a connected air conditioning system, a connected air purifier, smart watering system, and probably the most endearing of all, the self-driving connected car.
Given the number of new smart appliances that are being introduced as of late, it won’t be long before your home has not a few connected devices, but tens or even hundreds of gadgets, appliances and accessories that are all monitoring, generating data and communicating among themselves. And that’s just one household. Just how many families are in your block, neighborhood and town? The number of connected devices will grow exponentially as we get ‘smartified’.
How are the networks and data centers of today going to cope with the data explosion of tomorrow? That’s where innovations in technology such as Silicon Photonics, Knights Mill Xeon Phi, 5G, distributed computing and more, fall in place to bridge our smart and connected future!
Here are highlights from today’s keynote:-
To discuss more about 5G and its potential, Dr.Murthy Renduchintala called on other partners who’ve a vested interest in the 5G network for a roundtable discussion on stage:- Tom Keathley (SVP Wireless Network Architecture & Design for AT&T), Siezo Onoe (CTO of Docomo NTT) and John Gordon (CDO for Current powered by GE). Here are the key takeaways:-
Split second decisions to be made in critical situations. network architecture has to evolve so that edge devices can match servers. #IDF16— HardwareZone (@hardwarezone) August 17, 2016
LTE wasn't designed to handle 50 billion connected devices, but 5G will. The interconnectedness of everything is coming in 2018. #IDF16— HardwareZone (@hardwarezone) August 17, 2016
Here are some interesting anecdotes form the keynote:-
GPU architectures scale poorly on a complex deep learning model, especially with the lack of memory - CEO of Indico. #IDF16— HardwareZone (@hardwarezone) August 17, 2016
Training neural networks faster is the reason why Baidu has chosen Intel Xeon over other platforms to better A.I. #IDF16— HardwareZone (@hardwarezone) August 17, 2016
For more context on the announced Xeon Phi, check out our news piece.
Last, but not least, here’s the full keynote if you want to catch the full details:-