Device Big Bang: The Need for an Infrastructure
The novelty of the e-era (e as in e-government) was to elevate digital documents to legally binding status. Before that time a word processor was a tool for preparing paper documents.
The u-era (u as in ubiquitous) introduces a new breed of communicating participants: objects. The e-era was concerned with digital people-to-people communication. The u-era adds people-to-object and object-to-object channels. The number of communicating devices keeps growing at an explosive rate. The “device big bang” is here. Devices will outnumber people on the Internet, creating the Internet of things.
Let me suggest that a new infrastructure is needed. Why would devices need more infrastructure than just the Internet? Let’s go through an example.
Assume we upgrade your refrigerator with a digital camera, a remote temperature control chip and wireless network communication. We assign it an IP address and connect it to the Internet. What happens?
Security: Hordes of computers around the globe do nothing but connect to unrelated computers to seek for vulnerabilities in order to take control over them. Your refrigerator could be hacked in five minutes.
Aggregation: Some kind of overview is needed when the number of u-gadgets grows. For instance, an alert panel would be useful. Such aggregation must exist outside individual appliances.
Compatibility: Imagine your u-home limits your choice to only brand X refrigerators. We definitely want the network to absorb differences between appliances.
Presentation Layer: Good architectural taste urges us to keep number crunching apart from generating flashy graphics. It’s enough for a u-refrigerator to communicate digits. Presentation should be offloaded to a separate service.
Storage Layer: Time travel is very useful for discovering trends. Past data should be stored somewhere. For reasons similar to the presentation layer, such storage should be offloaded to a service outside appliances.
Supervision: The whole digital setup needs frequent health checks. If any of your u-home sensors stops working you would like to know it. Even if it says “temperature ok” one has to check that it is really alive.
The u-home provides an example of the things we expect to populate the Internet in vast numbers. Many different applications have to deal with the same issues, like security and offloading functionality to keep sensor cost down.
A subsequent blog post will outline a more specific architecture.
P.S. If the “ubiquitous” terminology is new to you, please see the reports in the Downloads section.