Network Intensive and Compute Intensive Hardware Acceleration: Difference between revisions

From NFVwiki
No edit summary
Line 1: Line 1:
== PoC Team ==
== PoC Team ==
BT - Huawei - EZChip - AMD - Tilera
BT  
 
Huawei
 
EZChip
 
AMD
 
Tilera


== Main Contact ==
== Main Contact ==

Revision as of 14:48, 22 July 2014

PoC Team

BT

Huawei

EZChip

AMD

Tilera

Main Contact

Evelyne Roch, Huawei evelyne.roch@huawei.com

PoC Demo

Stage 1: Video URL will be made available in August 2014

Stage 2: SDN and OpenFlow World Congress, 14-17 October 2014

Stage 3: Industry event in 1Q 2015 to be confirmed

Abstract

The main purposes of this PoC are to demonstrate the benefits of Hardware Acceleration in NFV environments and to identify detailed requirements for a Hardware Abstraction Layer to enable portability across different Hardware Acceleration platforms.

The benefits of Hardware Acceleration include an increase in performance, better resource utilization and cost reduction for some Network Functions. This PoC accelerates the following functions as examples of functions that can benefit from Hardware Acceleration: Load Balancing, IKE (Internet Key Exchange), Encryption and Video Transcoding.

This PoC consists of multiple stages demonstrating increasing levels of complexity of integration of Hardware Acceleration solutions from different vendors demonstrating Dynamic Optimization of Packet Flow Routing for Network Intensive Functions to dynamic Service Chaining. The PoC can form the basis for future Open Source and Standardization work in the definition of a Hardware Abstraction Layer.

PoC Proposal

NFVPER(14)000078_PoC_Proposal_-_Network_Intensive_and_Compute_Intensive_Hardw.docx

UPDATED proposal (08/07/2014): NFVPER(14)000078r1_PoC_Proposal_-_Network_Intensive_and_Compute_Intensive_Hardw.docx

Download PDF version