Simulcastr is the first thing IBM has delivered out of a vision of the future of technology that is emerging within IBM Research, Desai said. That vision has been influenced by the notion of open source software, where software developers all share and contribute to projects for the common good of all developers.
Today, about 90 percent of the digital data in the world is generated at the outer edge of networks -- via smartphones, sensors and the like, Desai said. The volume of this data is exploding and becoming harder to consume. “So, rather than handling it all via the Internet and cloud computing centers, why not use local WiFi networks and nearby computing and storage resources instead?” Desai said. “That way we can unlock the value of all that data.”
As the data explosion continues, there appear to be no physical limits to its growth. “One of the underlying reasons for this data growth is all the sensors and devices and wearables and machines and cars that are all on the ground sensing all kinds of phenomena, including temperature and accelerometer data and videos and audio are transmitting all of this sensor data to the cloud,” Desai said in the video. “These devices and vehicles and sensors that we’re talking about, bring with them their own connectivity, their own storage, their own computing power, their own sensors, and the ability to do something useful with this data right at the edge.”
A new approach to computing must emerge, he said. And this new approach to computing is going to harness together the power of connected devices.
“With our new approach to computing, people will instead share hardware resources--everything from WiFi hotspots to sensors to the underutilized storage and data processing capabilities on our smartphones and tablets,” Desai said.
However, this approach is not without its challenges. For instance, when connecting and managing a large collection of different kinds of devices that temporarily reside in a specific place, “the management system has to be able to identify the devices that are present, confirm that they’re ready to participate in peer-to-peer sharing, and then distribute computing tasks to the various devices in a highly-efficient way. Then the system has to aggregate the content from different sources, present viewing choices to participants, and serve up pieces of content to the people who choose to see them.”
This will require cognitive technologies such as image analytics that recognize specific people and activities in video streams, and deep learning algorithms that enable computers to gain knowledge through their interactions with data, Desai said.
However, when it is all said and done, Desai said he envisions several real-life scenarios where the Simulcastr technology could be used to benefit consumers, businesses, governments and other organizations. For instance, media companies could use it to gather real time video reports during a storm, a fire, a riot or a crime in progress. Or, insurance companies or police could use videos captured by connected cars to quickly gather evidence about the cause of a traffic accident, he said in his post.