By Janice J. Heiss on Sep 24, 2013
Jai Suri, Group Product Manager, Java Platform, at Oracle, with
Oracle Embedded Java architect, Noel Poore, gave a session that was both
practical and visionary, titled “Internet of Things with Java” that
provided a glimpse of the challenges and prospects faced by the coming
Internet of Things (IoT). Suri was quick to point out that the session
was not a showcase of Oracle solutions for IoT; nor would it provide
best practices or design patterns for IoT. “It’s too early and we’re not
there yet,” said Suri.
He pointed out that the potential range of IoT applications is vast, from home and industrial automation to improved healthcare. His team has been spending a lot of time and effort trying to figure out how Java fits into the IoT space.
The IoT market is relatively new and evolving, and full of proprietary technologies with, as yet, no standardization. The biggest challenge Java faces is creating a horizontal technology stack that addresses a wide range of needs and challenges. IoT, he pointed out is nothing new – machines connecting with other machines go back to the dawn of computing. But recently, new technologies have made it more accessible than ever before. In the US, Comcast offers XFINITY home automation offering remote monitoring, temperature, lighting and small appliance control, real-time alerts when doors or windows are opened, and more. In Europe, Deutsche Telekom offers a similar system.
In health care, remote patient monitoring is an area of rapid growth. IoT is making a difference in industrial automation and business optimization and efficiency. Other segments of IoT growth include building management, energy, consumer, retail, IT and networks. Research groups are predicting a market of somewhere around $350 Billion by 2017, some of which will be committed to technology.
According to Suri, various factors are driving the growth in IoT. First, connected devices are growing rapidly, with shipments expected to range from $50B to $200B by 2020. Moore’s law is allowing devices to become smarter and, as a result, connectivity is cheaper. Extra bandwidth is available to be redeployed for data traffic and channels are being created that allow companies to move data more cheaply.
IoT traffic data is being stored in databases to be analyzed, so data is growing rapidly. Business opportunities are increasing as the number of devices connected to the cloud allowing for the tracking of shipments, cars and other things grows.
Why is this different from a simple client and server? First the number of web, desktop and mobile applications talking to a server is rapidly increasing. Devices may be on batteries, Wifi, Bluetooth or a long range network – the complexity is huge. And most of these devices do not have a human operating them.
IoT lacks any standard protocol for communication among devices. Protocols depend upon the industry. In home automation, Bluetooth is common along with short range radio networks. Ultimately it is about how users get and receive data from devices. Suri pointed out that most developers give little thought to security. When he joined Oracle he received a badge that allows him to access and send protected data; the badge provides his identity, which governs access. But how do we put an identity on a temperature sensor connected to our home gateway? Or on data about our medical condition?
Some companies are building data centers that allow companies to connect their enterprise applications to a data center so they won’t have to worry about scaling. But is this the right approach when a company has invested millions of dollars in enterprise infrastructure? Why leverage what you already have?
Suri summarized the critical issues:
--Communication across multiple-protocol networks
--Software provisioning & lifecycle management across diverse devices
--Data acquisition from thousands of diverse devices
--Managing large volumes of fast data in a scalable architecture
--Leveraging existing enterprise architectures for evolving IoT needs
With most IoT solutions currently being written from scratch, the need for a horizontal platform seems obvious.
How Java Fits In
Noel Poore then showed how Java is the best fit for IoT, emphasizing that his focus was on the IoT with Java and not the Java Internet of Things. In a situation so fragmented with different device drivers, chip sets, operating systems and so on, the availability of a platform that allows developers to move code around with little worry about which device is running it is ideal.
Managing 50 billion devices sensibly and scalably constitutes a huge challenge. Poore presented a conceptual architecture with the pieces that need to be in place for a horizontal IoT platform to work. This would enable developers to build solutions based on platform rather than rebuilding the solution every time a different IoT problem must be solved. The conceptual architecture begins with wireless and wired sensors feeding into an initial gateway which feeds into a core network; in addition smart sensors may bypass the initial gateway and go directly to the network. The network feeds into an IoT communication gateway, which in turn feeds access management, IoT management and data routing and analysis. The first of these two feed into identity access and management, while data routing and analysis is sent to enterprise business and business intelligence to attempt to gain value from the data.
Suri closed by summarizing the take-home points of the session:
* IoT technologies are a “Wild West” full of proprietary implementations and a highly fragmented vendor ecosystem.
* Java enables an open and standards-based secure IoT platform that seamlessly integrates devices with enterprise applications.
* But significant innovations are needed across the platform and the ecosystem products to make this vision a reality.
* The good news is that Java is ahead of the curve, and very well positioned to become the de facto platform for IoT applications.
Look for podcasts of JavaOne sessions at Parleys.com starting in early October.