Accurate Identification of All Network Traffic

Accurate traffic classification forms the foundation of the value you get with Allot multiservice solutions. The ability to granularly identify and measure the numerous and diverse traffic flows determines the level to which network operators can quantify and understand network usage, optimize network efficiency, assure high-quality service delivery, create value added services and charge for their use.

Multi-Dimensional Traffic Classification

Our superior Traffic Classification is based on Dynamic Actionable Recognition Technology (DART) which employs multiple inspection methods to identify traffic according to Layer-7 application, user, device, access, video and contextual attributes. From deep packet inspection to heuristic analysis of IP flow behavior, Allot’s synergy of inspection methods results in highly granular and accurate recognition even at maximum speeds and peak loads.

Encrypted Traffic Classification

The traffic encryption that is widely used by Internet services and data privacy applications poses a serious challenge to traffic visibility in service provider and enterprise networks. Encryption guarantees the security and privacy of the user by concealing the data flow and preventing its identification. For example, P2P and VoIP services such as BitTorrent and Skype, obfuscate their protocol structure to avoid detection, traffic shaping (i.e., throttling), or metering. Likewise, many OTT applications use SSL to encrypt communications and hide user information.

Allot Dynamic Actionable Recognition Technology proactively learns and adapts to these changing tactics using:

  • Analysis of pattern matching of packet contents.
  • Analysis of numerical properties of one packet or several packets in flow
  • Heuristic analysis of behavior statistics extracted from inspected transactions.
  • Peer learning: analysis of peer system behavior to identify P2P seeders (multiple transmission of files) and popular peers
  • Port, IP address, or range of IP addresses classification
  • Machine learning based on statistical distribution patterns for over 1000 traffic attributes
  • Automatic IP discovery through analysis of service hosts and subscriber records
  • Predictive DPI (PDPI) is Allot’s patented technology that enables the classification engine to learn from the patterns and behavior of traffic that it recognizes with 98% accuracy, and compare those patterns and behaviors to encrypted flows in order to improve identification accuracy.

Hitless Updates keep you current

Frequent and hitless updates keep the traffic classification engine inside every Allot multiservice platform up-to-date, ensuring that you have full visibility of all the latest Internet applications and protocols. Hitless signature updates are propagated automatically to all platforms without affecting surrounding services or systems.