By Bob Eastwood
If Deep Packet Inspection (DPI) were a thing then I would dub it Lazarus, the chameleon-cat of technologies. Why? Like the biblical character Lazarus it is resurrected and, extending that metaphor, like a chameleon-cat it not only has nine lives but also the ability to change its color, name and functionality with each new resurrection. Being the oldest of the three technologies I’ve explored in blogs here at Engage including Big Data and SDN, DPI landed on the back side of Gartner’s hype cycle on the “Plateau of Productivity” and enjoyed a similar if less inflated hype cycle well over a decade ago. It promised to give network operators and administrators visibility into, and control over, what was happening on their networks. At the time, traffic visibility was about as clear as mud. During the 90's, DPI was initially developed and used with success in applications like firewalls, detection of network abuse, and rudimentary statistical analysis. However, the granular level of visibility and control it enables has proven so effective, in-demand and mutable that it continues to find new applications to this day meaning that Gartner’s “Plateau” for DPI would demonstrate a continuing growth slope up and to the right.
Unlike Shallow Packet Inspection which refers to the inspection of packet headers for optimization of traffic routing, DPI is a form of network traffic filtering that examines not only the packet header, but also the content of the data packets as they pass through an inspection point on the Internet’s infrastructure. The characteristics of the packets are referred to as signatures and can be categorized into groups to enable a variety of uses. These signatures are now numerous and sophisticated enough to provide awareness of characteristics such as subscriber ID and service plan, bandwidth usage, device type, location, application type, content categorization, network conditions, traffic management status, and routing and peering attributes among others. The most advanced of these technologies now provides this information in real-time, allowing users to drill down into these categories for monitoring, troubleshooting, reporting, new service creation and analytics.
While older and less popular with the media, DPI is no less important a technology for network operators, especially for its latest role assisting increasingly capacity-stressed mobile operators to optimize and monetize their networks and in its latest relationship as an enabler for the two other current tech celebrities, Big Data and SDN. Infonetics found that spending on standalone DPI solutions grew 30% year-over-year in 2013, driven largely by the mobile segment, and that it was on track to reach nearly $2B by 2017. DPI has become a strategic weapon for mobile operators as they seek to control traffic in increasingly congested networks and to monetize data by offering personalized service plans. It is especially important for keeping track of offerings like family plans that may see multiple mobile devices using network bandwidth in different geographic areas, and tracking that usage against one service plan or data cap.
In the 2000s, DPI found a new life, “color” and name and with a solution it enabled called policy enforcement, whereby operators could take the signatures and apply policy rules to manage and control network traffic for infrastructure optimization to reduce costs, to better plan CapEx, and for the creation and marketing of new usage-based services for improved monetization of network infrastructure investments. DPI now enables broadband offerings such as service tiers, shared data plans, and applications such as bandwidth on demand among others. As shown in the accompanying chart, these new policy enforcement applications are proving very important to operators.
So, what’s next for DPI? All the while capable of providing some level of visibility and monitoring, the relentless march of Moore’s law and its concomitant rate of improvement in processing speeds and the storage capabilities of databases are leading to DPI’s next application—providing real-time network intelligence through telecom analytics. By teaming up with its newest companions among Gartner’s hype cycle alumni, SDN and Big Data, DPI promises to live on in ever-mutable ways. The story of Lazarus the chameleon-cat is to be continued.