Smart objects have the potential to bring the same visibility and intelligence to the operation of non-corporate environments such as military bases, towns and cities. IBM has developed the software that monitors a network of smart objects that automate the location, deployment and management of a city's physical assets. Using IBM's Maximo Asset Management software, a town or city could process the data from smart objects embedded within tens of thousands of physical assets such as water mains, traffic lights, bridges, park lawns, fire hydrants, garbage trucks and storm water ditches to provide city officials with comprehensive real-time pictures of its infrastructure.
In addition to providing quicker response to service outages and other emergencies, such a system could allow city officials to correlate data from multiple sources to spot hidden problems. One recent example of this capability is the Smart City Asset Management project in Corpus Christi, Texas, where analysis of the data from 3,843 water main breaks occurring over three years revealed that most of the failures originated in smaller-sized water mains. As a result, the town decided to minimize future failures by upgrading the trouble-prone mains with larger pipes.
Fletcher expects that, eventually, virtually everything within the enterprise environment will be connected. He cautions, however, that this will be a gradual evolution as today's proprietary systems give way to ones that use IP-based protocols which can easily share a common data infrastructure and management interface.
Integrating legacy systems usually involves some sort of bridging element used to translate digital protocols and electrical signals. For example, a software shim layer can be used to bridge the proprietary protocols used in many building automation systems to make them visible to an enterprise's IP-based management tools. While bridging is necessary for an application that is being built upon a large installed base of proprietary networks, it always adds an extra layer of cost and complexity to a solution.
During this transition period, legacy system elements will pose a second challenge because many of them lack any intelligence at all. In these cases, the objects will use the intelligence within the network to create what Fletcher calls "proxy intelligence." At its simplest, proxy intelligence is exemplified by "dumb" RFID tags which can be read by smart readers, making it possible for a central data infrastructure to track packages, equipment and other assets as they travel throughout the enterprise.
Likewise, most servers, routers and other IT equipment used in today's corporate LANs do not posses smart objects' self-awareness and self-reporting capabilities. In this case, management software can probe network endpoints, use their behavioral characteristics to positively identify most of them, and to make (very) educated guesses about the rest.
Security issues
For all the pain points a smart object-enabled enterprise relieves, the technologies that underpin its operation can be a pain point themselves because of the security issues they create. Kurt Stammberger, CISSP, vice president of market development at Mocana, says that extending an enterprise network beyond traditional MIS boundaries creates many additional external interfaces that inevitably invite attempts at tampering, hacking and other mischief.
Adding to the problem is the IP protocol itself. The fact that it is a well-known open standard makes it an excellent common system interface, but it also means that it does not have the (admittedly small) fig leaf of "security by obscurity" enjoyed by proprietary protocols like LonWorks or BACnet.
Stammberger adds that deliberate hacker attacks are only half of the security challenge. "There is enough random malware on the net that collateral damage from random, unintentional attacks is now not a possibility but almost a certainty," he says. "Polymorphic malware and swarms of botnets are now so prevalent and adaptive that an attack on a nearby wireless access point could result in collateral damage in unintended targets, such as disrupting a building's HVAC system or, worse yet, crashing a piece of networked medical equipment."
Protecting smart object-based systems in the Internet of things differs from traditional security models in several significant ways. Unlike the PC/server world, the embedded systems that support smart objects can be built from thousands of combinations of well over 50 commercially available CPU architectures and dozens of operating systems, each of which is likely to have multiple distinct application-specific variants.
"The result is a digital environment that looks much more like a rainforest than the PC world, whose limited choice of processor architectures and OSs resembles the endless rows of one variety of corn you see in the mono-crop farms of the Midwest," Stammberger says. One challenge posed by this "digital rainforest environment" is the wide range of computing capabilities found in smart objects, which range from powerful smartphones to 8-bit embedded sensors.
Security technologists are addressing these diverse requirements with software architectures that are more flexible and scalable than those used in PC-based applications. One of the most common approaches today is the modular software architecture used by Mocana which can be scaled in a granular manner to suit a particular platform and application. In a similar manner, the software's cryptography capabilities must be scalable to accommodate a particular platform's limited compute resources or to support enhancement with new, application-specific features such as IPSec, SSL, SSH, digital certificates and support for embedded VPN clients.
It's impossible to tell what new threats will emerge or which architecture will be the best solution as smart object technology continues to evolve, but there is at least one thing that developers can count on: Enterprise computing has entered an age where we will worry less about the cost of security than the cost of not having it.
The short case studies we've seen here are good examples of how smart object technologies have already begun to revolutionize the way businesses, institutions and governments operate. Their ability to unify previously fragmented data resources and provide visibility into parts of their operation that were not possible before can improve the efficiency of nearly any internal process, tighten supply chains and enable delivery of many new and innovative services to an organization's clients and customers. In addition to enabling increased profitability, the efficient use of resources made possible by smart object-enabled systems will play a key role in helping organizations of all kinds reduce their environmental footprints. They will also play a similar role in "smart cities" and other smart government initiatives.
The long-term success of all these applications will in large part depend on the widespread adoption of a standards-based approach that includes the IP protocol as part of the foundation of smart object technologies. The IPSO Alliance has emerged as the thought leader driving broad smart object interoperability with IP as the common networking protocol.
IP-based communication protocols will make it possible to build smart object systems using inexpensive, off-the-shelf silicon and software. IP-enabled smart object products will also enjoy better market acceptance, thanks to their ability to use existing private and public networks as a bridge between the closed MIS environment and real world applications. Developers and users alike will benefit from the rich ecosystem of interoperable products and innovative services made possible by the well-understood protocols that power IP-based smart object applications.