Home > Communication Service Provider Solutions > Telecom Multicloud Foundation > Red Hat > Guides > Red Hat Open Stack Platform Guides > Jumbo frames Performance Analysis Over Red Hat OpenStack 16.1 Infrastructure > Introduction
Software-Defined Network (SDN) and network function virtualization (NFV) have been driving the network transformation for many years, resulting in the majority of networking solutions virtualized. There are still challenges in getting optimum performance for services running on underlying virtual infrastructures.
Communications Service Providers (CSPs) support many services, for example, broadband access services, Voice over IP (VoIP) services, call center, IPTV and cable services. One of the most important measurements is the network capability to support Zero Packet Loss (ZPL). Any packet loss means disruption to services, especially for emergency call centers. For example, a 911 call drop is unacceptable. Traditionally, fixed network appliances can deliver the performance needed, and they do so with assurance. With SDN and NFV transformation, the Zero-Packet Loss capability to show expected performance over the infrastructure becomes extremely important.
Various laboratory measurement observations are not all published. Part of the problem is knowing whether the measured performance is, in fact, the expected performance. Typically, the DPDK open-source community published performance data for Physical Function and VirtIO for the Network Interface[1]. This is a great reference point to baseline expected performance. However, the DPDK open-source community doesn’t cover the test cases for jumbo packet size[2].
This collaborative whitepaper between Dell and Intel aims to provide guidance on jumbo packet size for Zero Packet Loss (ZPL) test cases over Dell EMC Ready Architecture (RA) Red Hat OpenStack Infrastructure. In addition, Red Hat engineering team also helped in finalizing result section and verified the best practices applied in this activity. The key learnings and best-known practices documented in this paper can help your testing and analysis.