PeerSec Networks
Enterprise Level Security for Devices

MatrixSSL Telecom / Datacenter Solutions

PeerSec Networks MatrixSSL meets memory usage and high availability requirements for servers handling thousands of secure connections.

Datacenter Security Highlights

  • Increasing device security places heavy requirements on servers
  • Telecom servers must often handle over 10,000 simultaneous secure connections.
  • MatrixSSL allows 10 times the active SSL sessions provided by other alternatives
  • Reduces failover recovery times and associated memory hardware costs
  • MatrixSSL is standards based and interoperable with clients including all major Web browsers, OpenSSL and GnuTLS
  • SSL/TLS support for standard datacenter operating systems including Microsoft Windows and Linux

Scaling to Thousands of SSL Sessions

Telecom providers must often scale to over 10,000 simultaneous calls. As secure VoIP calls become standard, this places a large burden on the server hardware. The CPU can offload much of the computation to hardware cryptography accelerators, but total memory usage becomes a limiting factor.

High Availability Requirements

If a failure occurs on a server handling thousands of sessions, the connections should fail-over to a backup to avoid dropped connections. Replicating the state of thousands of SSL connections is time consuming and costly. Minimizing the dynamic memory usage of SSL sessions increases the failover recovery speed, minimizes replication times and reduces the need for costly hardware replication solutions.

Platform Requirements

MatrixSSL implements standards based SSL and TLS protocols in a compact library. The memory requirements per SSL session are 12KB during initial handshake and under 4KB after session establishment. The small dynamic memory footprint of MatrixSSL allows 10 times the number of active connections in the same memory space as do other common SSL implementations. MatrixSSL's cryptography layer supports plug-ins for various hardware cryptography accellerators, providing a solution that is low in both CPU and memory usage.