Data Plane Development Kit: Difference between revisions

Revision as of 07:56, 21 June 2021

Revision as of 13:46, 21 June 2021

Line 28:

The EAL also provides additional services including time references, generic [[Bus (computing)|bus]] access, trace and debug functions and alarm operations.

Using DPDK libraries one can implement a low overhead [[run-to-completion]], pipeline or staged, event driven, or [[hubridhybrid]] model completely in userspace eliminating kernel and kernel to user copy. Hardware assists from [[NIC/Regex/Accelerators]], libraries enhanced to make use of Intelligence Storage Acceleration (ISA)<ref>{{Cite web|title=How DPDK can Deflate your Data! {{!}} SNIA|url=https://www.snia.org/educational-library/how-dpdk-can-deflate-your-data-2018|access-date=2021-06-21|website=www.snia.org}}</ref> for bulk performance and accessing devices via [[Polling (computer science)|polling]] helps to eliminate the performance overhead of [[interrupt]] too. Hugepages are used for large memory pool allocation, to decrease the amount of lookups and page management.<ref>{{Cite web|last=PANTHEON.tech|date=2020-01-09|title=VPP 105: Memory Management & DPDK APIs|url=https://pantheon.tech/vector-packet-dpdk-apis/|access-date=2021-06-21|website=PANTHEON.tech|language=en-US}}</ref>

The DPDK also includes software examples that highlight best practices for software architecture, tips for data structure design and storage, application profiling and performance tuning utilities and tips that address common network performance deficits.