Flexible Workload Acceleration on Intel® Architecture
Intel® QuickAssist Technology eliminates the need for various add-on acceleration modules
There will always be classes of workloads and algorithms that push the capabilities of traditional compute platforms. A prime example is computationally-expensive public key encryption algorithms supporting SSL, particularly the handshaking mechanism that allows a server and a client to establish a connection. Consequently, IT professionals often deploy specialty based add-on acceleration modules to increase performance and avoid overloading their servers.
Now, there’s an alternative to using specialized ASICs for compute-intensive workloads. Servers, based on the next generation communications platform from Intel, codename Crystal Forest, have new capabilities for accelerating asymmetric and symmetric cryptography, data compression and other workloads. The platform incorporates Intel® QuickAssist Technology, which includes acceleration modules that are accessed via a unified set of industry-standard application programming interfaces (APIs). They provide consistent conventions and semantics across multiple accelerator implementations, thus future-proofing software investments. The platform also delivers exceptional packet processing performance by running performance-optimized libraries from the Intel® Data Plane Development Kit (Intel® DPDK).
This paper provides an overview of Intel QuickAssist Technology and describes four implementation scenarios that underscore the flexibility available to equipment manufacturers, independent software vendors (ISVs) and IT professionals. Using this workload acceleration technology, it’s possible for Telco central offices and datacenters to significantly lower their equipment costs, and improve workload balancing across servers.
Read the full Flexible Workload Acceleration on Intel® Architecture white paper: http://www.intel.com/content/www/us/en/communications/communications-quick-assist-paper.html