Network Computing is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

Teak, Neterion Demo Solution

SAN FRANCISCO -- Blade.org members
Teak Technologies and Neterion are demonstrating their 10-Gb
Congestion-Free Ethernet switching solution for the IBM BladeCenter
platform at the NGDC Conference and Expo here today.

The demonstration features an IBM BladeCenter-H chassis, two Teak
10-Gb CFE Switch Modules, and six Teak 10-Gb CFE Expansion Cards
running on a mix of six AMD/Intel LS-21/HS-21 server blades. The two
switch modules simulate a two-hop network. A modified ping-pong test
running across multiple NetPIPE clients/servers simulates network
congestion and displays application-level results graphically to
highlight performance improvements under congestion. The results
demonstrate in excess of a ten-fold improvement in end-to-end
application-level latency, fairness, and full link utilization across
multiple flows under transient and sustained network congestion.

The Teak 10-Gb CFE Expansion Card for the IBM BladeCenter integrates
Neterion's advanced ASIC featured in the company's family of 10-GbE
Xframe(R) V-NIC(TM) adapters featuring the Hyperframe(TM) suite of
technologies designed to provide virtualization at the I/O level. The
ASIC is based on a sophisticated state machine architecture that
delivers the highest performance and reliability available, and as a
result, enables complete end-to-end virtualized networks.

Teak's 10-Gb CFE switching solutions unlock the full promise of a
loss-free Ethernet thereby transforming a data center network into a
Congestion-Free Zone, and enabling performance that:

  • improves upon comparable best-effort 10-GbE price/performance by up to
    16x
  • maximizes load-invariant, loss-free link utilization up to 100 percent
  • reduces number of uplinks in multi-hop network configurations by up to
    4x
  • delivers SLA-driven, predictable application performance
  • 1