The “Cloud of Things” builds on the “Internet of Things” (a term coined at the MIT Auto-ID Labs), where information about objects is accessed via the Internet; and machine-to-machine (M2M) computing, where wireless communication protocols enable peer-to-peer exchange of data between electronic devices. Constructing a model of an object in the cloud with a defined set of Application Programming Interfaces (APIs) facilitates the integration of data from heterogeneous sources more readily than trying to establish a common registry or protocol across multiple organizations.
The “Cloud of Things” initiative will bring together researchers and industry to design sponsored-research initiatives for specific companies and industries; it will also host a series of theme-focused workshops, conferences, panels, demonstrations, exhibits and pilots on various topics. Particular areas of focus will include manufacturing, retail, health care, supply chain and more. The initiative is open to companies, nonprofits and individuals interested in promoting the development, adoption and commercial success of big data applications.
The concept of a “Cloud of Things” as an organizing principle for cloud computing was announced by Department of Mechanical Engineering Professor Sanjay Sarma, the co-founder and research director of the Auto-ID Labs, during the Auto-ID Labs Big Data Conference and Startup Challenge, held Oct. 9-10 at Bartos Theater. The event convened academic and industry thought leaders from cloud computing, network science and mobile/social retail commerce. Presenters discussed opportunities for aggregating data to enhance businesses performance — from sources as diverse as city infrastructure, vehicles and buildings — with structured data from traditional business applications.
The event was co-hosted with GS1-EPCglobal, Voluntary InterCommerce Solutions (VICS), the Mass TLC Big Data Cluster and the MIT Industrial Liaison Program.
In his presentation at the event, Mike Olson, CEO of Cloudera, an open-source big-data software-development company, discussed how businesses are struggling with how to analyze the 90 percent of business data that is not in traditional relational structures. He cited the 104 petabytes of information that Facebook currently supports as a testament to the resilience of open-source Hadoop file sharing and Map Reduce distributed-computing infrastructure. In his keynote later that day at a dinner reception at the MIT Museum, Mike Stonebraker, a member of the MIT Computer Science and Artificial Intelligence Laboratory’s big-data initiatives, noted that companies with interest in analyzing big data to offer real-time business benefits to customers are fast headed into a wall, depending on the characteristics of their data, independent of whether one has joined No SQL or New SQL movements.
In the discussions of “hot-button” issues in implementing big data for omni-channel and mobile retail commerce, the panel pointed to an urgent need for industry best practices to provide context for granular event data being generated by item-level serialized barcode and radio-frequency identification (RFID) scans in retail environments today, where traditional scanners are used alongside smartphones. A specific example of what is needed was given in an address by Abhi Dhar, chief technology officer of Walgreens’ e-commerce business, who pointed to the need for specifications and best practices for storing, monitoring and analyzing medical-device data.
Some organizations and business that are already supporting the initiative include Walgreens, BJ’s Wholesale Club and the Mass Technology Leadership Council Big Data Cluster.
“Our goal is to stimulate the creation of innovation through the rapid development of big-data applications based around APIs for objects and infrastructure from the physical world that can be accessed on the Web via smartphones and other electronic devices,” said Steve Miles, the conference organizer and an MIT research scientist who is currently heading big-data business development at Auto-ID Labs. “There are many Boston-area individuals and organizations in the big-data ecosystem — including both established companies and startups — that have critical assets that can help make a ‘Cloud of Things’ successful.”