What is Deep Packet Inspection (DPI) ?
Deep Packet Inspection (DPI) software analyzes a data packet as it passes an inspection point in order to determine the transported protocol and/or application and other metadata of the traffic. This requires maintaining a software library and understanding patterns and behaviors for every new app and version otherwise a packet can easily be incorrectly classified. Advanced Deep Packet Inspection (DPI) technologies, behavioral, heuristic and statistical analysis are used to accurately detect applications in real-time, including popular encrypted applications, e.g. WhatsApp, Skype, BitTorrent, eDonkey, etc. DPI can extract metadata and application attributes from the network traffic to determine application performance and quality of service for VoIP and video, as well as to protect against network security threats.
DPI software is made to inspect packets at high wire speeds while maintaining high standards of accuracy and performance for application classification and metadata extraction. Since the software is typically integrated in the network and security infrastructure, it is critical to keep the resource requirements at highly optimized levels. Optimization of the resource requirement such as the number of cores (on a multi-core processor) and the amount of on-board memory needed by the software engine is thus highly recommended for the integrated DPI and application classification technology.
DPI Drivers?
Growing demand for DPI is coming from mobile networks. New mobile devices, from tablets and smartphones to even wearable devices are emerging every day along with new applications and services with widely varying traffic patterns and bandwidth needs. DPI is used in mobile networks for traffic management and policy control, performance monitoring and troubleshooting, prioritization for QoE, content caching and data offload purposes. Analytics is another area where operators are starting to use DPI to analyze data usage to enhance data packages and improve QoE per application e.g. video streaming services.
Ensuring that video bandwidth is properly managed while assuring high Quality of Experience for critical time sensitive services requires DPI solutions capable of identifying, measuring and analyzing these types of services at high-speed, under any conditions with high reliability.
In addition as mobile networks move to IP via LTE, it’s crucial for operators to ensure that VoLTE and other LTE services are delivered with a comparable QoS as those on the traditional circuit switched services. To deliver this, the operators need to both monitor and manage the QoS of the actual network and service, and measure the QoE for the users of the VoLTE application.
Mobile data offload solutions typically treat all traffic flows equally without distinguishing the application involved or the device used, yet solutions are starting to emerge that selectively steer traffic flows based on the application used and DPI. A new use case for DPI was demoed at Intel Developer Forum 2013 radically enhances Quality of Experience of delay sensitive apps.
DPI Trends
Network Analytics and optimisation: Network and customer insights derived from the DPI technology will become increasingly important in addressing the growing demand for bandwidth intensive applications such as video. According to Cisco’s Visual Networking Index, consumer Internet video traffic will make up over two-thirds (69%) of all consumer Internet traffic in 2017. When this figure is extended to include video traffic exchanged through peer-to-peer (P2P) file sharing, the share of video traffic will reach the range of 80-90% by 2017.
Internet of Things: DPI is likely to be a key part of Machine to Machine (M2M) strategies going forward, enabling identification and mitigation of the traffic generated by connected devices, and enabling QoS guarantees for services such as telematics and remote patient monitoring.
Devices: DPI is starting to move from the mobile core and access networks to devices themselves, enabling operators to better manage signaling traffic, provide mobile security, improve SLAs on enterprise mobility apps and enable more granular subscriber controls such as parental control and shared data plan device control for consumers.
“Article contributed by Ipoque – www.ipoque.com.”
