Algorithms, the cloud, the internet of things, and data — a combination referred to as ACID — are changing the face of computing, Microsoft corporate VP Joseph Sirosh said today at the Microsoft Ignite conference in Atlanta.
Sirosh showed numerous ACID offerings from both Microsoft and other firms during a session titled “The Unreasonable Effectiveness of ACID.” That title seemed to be a double pun: on the book “The Unreasonable Lightness of Being” and on the acronym “ACID.”
RELATED: Adobe partners with Microsoft to use Azure as ‘preferred solution’ for cloud services
The acronym for some 30 years has referred to one of the fundamental principles of relational databases, a longtime IT tool. It stands for “atomicity, consistency, isolation and durability,” the four theoretical properties of database transactions. Now relational databases, though still important, are being challenged to some extent by so-called NoSQL databases, more modern products that store data in looser form and don’t require the rigorous set-up and querying procedure of relational databases.
For example, Sirosh demonstrated Azure DocumentDB, a noSQL database that he said can be scaled instantly in multiple locations, making it ideal for use with mobile apps that access data. Putting data close to the user, wherever the user is located, cuts latency to single milliseconds, he said. In contrast, accessing data a continent away takes hundreds of milliseconds. DocumentDB can handle “hundreds of millions of requests per second,” he said.
But SQL databases weren’t ignored. In fact, they were at center stage.
The combination of the new SQL Server 2016 and Windows Server 2016 can put 24 terabytes of data into memory, for hugely improved performance compared to storing it even on a solid-state disk, Sirosh said. “It basically eliminates the need for storage and network,” he said.
Symitar, an independent software vendor with a service that predicts loan defaults, showed the use of SQL Server to analyze 20 million loans in 17 seconds, using the R statistical computing language. That’s really fast.
Sirosh showed how algorithms are deployed on a big-data processing engine, the Azure Data Lake repository and Power BI analytics to analyze the content of millions of images, giving each the relevant tag (“person,” “horse,” “dog”) and analyzing human faces for emotion.
He also showed a demo of those products analyzing the sentiment of paragraphs within “War and Peace,” plotting each of 50 characters’ prominence within each chapter and changes in their fortunes and emotional states throughout the book. “I could generate these guides for every page in every book in the Library of Congress, or for every call made into a call center,” he said.
As to sensors, they’re expected to increase 100-fold by 2020, Sirosh said: from 500 million in 2013 to 50 billion in 2020. Rolls-Royce is building them into jet engines for real-time monitoring, race cars are sending data to their drivers to optimize the timing of pit stops and football players’ shoulder pads are helping to track their movements and speed.
