Vedicis technology represents the new generation of Deep Packet Inspection (DPI) and Policy Enforcement garnering both high accuracy in usage traffic analysis and content policy enforcement. By eliminating current technical bottlenecks for memory access to analyse traffic in real time with rich contextual information, this breakthrough approach for DPI & Policy Enforcement 2.0 expands the scope of DPI technology and enables new features to manage broadband without latency in the network.
Its innovative software architecture provides a high level of flexibility for integration and/or new software modules deployment while achieving stunning performance for broadband traffic management.
Challenges for Policy Enforcement
To manage million policies in broadband traffic, that depend on multiple key network criteria, new Policy Control & Charging Enforcement systems require
- very fast access to RAM, to prevent any traffic latency because of processing,
- large RAM size, to take into account all criteria, including bandwidth calculation in real time, volume/quota follow up, protocols applications classification, ip address user and profile mapping.
Current platforms have been developed with specialized hardware including ASIC and NPU (Network Processor Unit) to design purpose-built networking cards. But these traditional router and switch architectures do not present the optimized access to RAM, with large enough RAM size, unless the number of specific cards is multiplied, representing additional complexities, for instance with internal load balancers …
New system with Intel Servers to end RAM access bottlenecks
Vedicis powerful software platform taps the full performance of Intel standard servers, either 2U – 4U or Blade systems, based upon Intel Quick Path Interconnect architecture for deployment on 1 & 10 Gbps optical links. In this server architecture, the direct traffic routes from the network adapter to the CPU and L3 cache then to the RAM have suppressed many traditional bottlenecks that are found with other card architectures based on network processors (NPU) or ASIC.
High performing platform to manage traffic
Current Vedicis performance gets a boost every 18 months thanks to Moore’s Law. Today’s platform with dual Hexacore Xeon system manages 10 Mega Packets per second per system with a latency below 30 microseconds.
Vedicis Software Platform: Linux flexibility and speed for integration
While traffic packets are managed at the Linux kernel level, applications and service logic are managed at the user level. This enables a critical flexibility to develop new APIs for integration on top of existing APIs for a fast integration with databases, charging and billing systems, radius and user provisioning gateway or policy servers.
Carrier grade platform for secured and scalable deployment
Vedicis offers various high availability scenarios, either optical bypass, or clusters with active / cold-standby or hot-standby to ensure service and traffic resiliency. Vedicis platform supports the redundant architectures that are already in place and based upon VRRP, RSTP and link aggregations.
The software architecture has been optimized to grow and scale with performance to manage multiple 1 GbE and 10 GbE aggregated links. It meets the wire line operator’s new standards for peering points and backbone/aggregation, and Mobile operator’s needs at the GGSN level.
Its centralized management with web GUI or command lines is designed for operation efficiency. Standard APIs are available for rapid and smooth integration within operators' management framework.