Benchmarking Virtual Switches in OPNFV

Benchmarking Methodology Working Group (BMWG) has traditionally conducted laboratory characterization of dedicated physical implementations of internetworking functions. The Black-box Benchmarks of Throughput, Latency, Forwarding Rates and others have served our industry for many years. Now, Network Function Virtualization (NFV) has the goal to transform how internetwork functions are implemented, and therefore has garnered much attention. This memo describes the progress of the Open Platform for NFV (OPNFV) project on virtual switch performance characterization, "VSWITCHPERF". This project intends to build on the current and completed work of the Benchmarking Methodology Working Group in IETF, by referencing existing literature. For example, currently the most referenced RFC is (which depends on ) and foundation of the benchmarking work in OPNFV is common and strong. See https://wiki.opnfv.org/characterize_vswitch_performance_for_telco_nfv_use_cases for more background, and the OPNFV website for general information: https://www.opnfv.org/ The authors note that OPNFV distinguishes itself from other open source compute and networking projects through its emphasis on existing "telco" services as opposed to cloud-computing. There are many ways in which telco requirements have different emphasis on performance dimensions when compared to cloud computing: support for and transfer of isochronous media streams is one example. Note also that the move to NFV Infrastructure has resulted in many new benchmarking initiatives across the industry, and the authors are currently doing their best to maintain alignment with many other projects, and this Internet Draft is evidence of the efforts.

The primary purpose and scope of the memo is to inform BMWG of work-in-progress that builds on the body of extensive literature and experience. Additionally, once the initial information conveyed here is received, this memo may be expanded to include more detail and commentary from both BMWG and OPNFV communities, under BMWG's chartered work to characterize the NFV Infrastructure (a virtual switch is an important aspect of that infrastructure).

This section highlights some specific considerations (from )related to Benchmarks for virtual switches. The OPNFV project is sharing its present view on these areas, as they develop their specifications in the Level Test Design (LTD) document.

To compare the performance of virtual designs and implementations with their physical counterparts, identical benchmarks are needed. BMWG has developed specifications for many network functions this memo re-uses existing benchmarks through references, and expands them during development of new methods. A key configuration aspect is the number of parallel cores required to achieve comparable performance with a given physical device, or whether some limit of scale was reached before the cores could achieve the comparable level. It's unlikely that the virtual switch will be the only application running on the SUT, so CPU utilization, Cache utilization, and Memory footprint should also be recorded for the virtual implementations of internetworking functions.

External observations remain essential as the basis for Benchmarks. Internal observations with fixed specification and interpretation will be provided in parallel to assist the development of operations procedures when the technology is deployed.

A key consideration when conducting any sort of benchmark is trying to ensure the consistency and repeatability of test results. When benchmarking the performance of a vSwitch there are many factors that can affect the consistency of results, one key factor is matching the various hardware and software details of the SUT. This section lists some of the many new parameters which this project believes are critical to report in order to achieve repeatability. Hardware details including: Platform details Processor details Memory information (type and size) Number of enabled cores Number of cores used for the test Number of physical NICs, as well as their details (manufacturer, versions, type and the PCI slot they are plugged into) NIC interrupt configuration BIOS version, release date and any configurations that were modified CPU microcode level Memory DIMM configurations (quad rank performance may not be the same as dual rank) in size, freq and slot locations PCI configuration parameters (payload size, early ack option...) Power management at all levels (ACPI sleep states, processor package, OS...) Software details including: OS parameters and behavior (text vs graphical no one typing at the console on one system) OS version (for host and VNF) Kernel version (for host and VNF) GRUB boot parameters (for host and VNF) Hypervisor details (Type and version) Selected vSwitch, version number or commit id used vSwitch launch command line if it has been parameterised Memory allocation to the vSwitch which NUMA node it is using, and how many memory channels DPDK or any other SW dependency version number or commit id used Memory allocation to a VM - if it's from Hugpages/elsewhere VM storage type: snapshot/independent persistent/independent non-persistent Number of VMs Number of Virtual NICs (vNICs), versions, type and driver Number of virtual CPUs and their core affinity on the host Number vNIC interrupt configuration Thread affinitization for the applications (including the vSwitch itself) on the host Details of Resource isolation, such as CPUs designated for Host/Kernel (isolcpu) and CPUs designated for specific processes (taskset). - Test duration. - Number of flows. Test Traffic Information: Traffic type - UDP, TCP, IMIX / Other Packet Sizes Deployment Scenario

Virtual switches group packets into flows by processing and matching particular packet or frame header information, or by matching packets based on the input ports. Thus a flow can be thought of a sequence of packets that have the same set of header field values or have arrived on the same port. Performance results can vary based on the parameters the vSwitch uses to match for a flow. The recommended flow classification parameters for any vSwitch performance tests are: the input port, the source IP address, the destination IP address and the ethernet protocol type field. It is essential to increase the flow timeout time on a vSwitch before conducting any performance tests that do not measure the flow setup time. Normally the first packet of a particular stream will install the flow in the virtual switch which adds an additional latency, subsequent packets of the same flow are not subject to this latency if the flow is already installed on the vSwitch.

This outline describes measurement of baseline with isolated resources at a high level, which is the intended approach at this time. Baselines: Optional: Benchmark platform forwarding capability without a vswitch or VNF for at least 72 hours (serves as a means of platform validation and a means to obtain the base performance for the platform in terms of its maximum forwarding rate and latency).

Benchmark platform forwarding capability Benchmark VNF forwarding capability with direct connectivity (vSwitch bypass, e.g., SR/IOV) for at least 72 hours (serves as a means of VNF validation and a means to obtain the base performance for the VNF in terms of its maximum forwarding rate and latency). The metrics gathered from this test will serve as a key comparison point for vSwitch bypass technologies performance and vSwitch performance.

Benchmark VNF forwarding capability Benchmarking with isolated resources alone, with other resources (both HW&SW) disabled Example, vSw and VM are SUT Benchmarking with isolated resources alone, leaving some resources unused Benchmark with isolated resources and all resources occupied Next Steps Limited sharing Production scenarios Stressful scenarios

The overall specification in preparation is referred to as a Level Test Design (LTD) document, which will contain a suite of performance tests. The base performance tests in the LTD are based on the pre-existing specifications developed by BMWG to test the performance of physical switches. These specifications include: Benchmarking Methodology for Network Interconnect Devices Benchmarking Methodology for LAN Switching Device Reset Characterization Packet Delay Variation Applicability Statement In addition to this, the LTD also re-uses the terminology defined by: Benchmarking Terminology for LAN Switching Devices Packet Delay Variation Applicability Statement Specifications to be included in future updates of the LTD include: Methodology for IP Multicast Benchmarking Packet Reordering Metrics As one might expect, the most fundamental internetworking characteristics of Throughput and Latency remain important when the switch is virtualized, and these benchmarks figure prominently in the specification. When considering characteristics important to "telco" network functions, we must begin to consider additional performance metrics. In this case, the project specifications have referenced metrics from the IETF IP Performance Metrics (IPPM) literature. This means that the test of Latency is replaced by measurement of a metric derived from IPPM's , where a set of statistical summaries will be provided (mean, max, min, etc.). Further metrics planned to be benchmarked include packet delay variation as defined by , reordering, burst behaviour, DUT availability, DUT capacity and packet loss in long term testing at Throughput level, where some low-level of background loss may be present and characterized. Tests have been (or will be) designed to collect the metrics below: Throughput Tests to measure the maximum forwarding rate (in frames per second or fps) and bit rate (in Mbps) for a constant load (as defined by RFC1242) without traffic loss. Packet and Frame Delay Distribution Tests to measure average, min and max packet and frame delay for constant loads. Packet Delay Tests to understand latency distribution for different packet sizes and over an extended test run to uncover outliers. Scalability Tests to understand how the virtual switch performs as the number of flows, active ports, complexity of the forwarding logic’s configuration… it has to deal with increases. Stream Performance Tests (TCP, UDP) to measure bulk data transfer performance, i.e. how fast systems can send and receive data through the switch. Control Path and Datapath Coupling Tests, to understand how closely coupled the datapath and the control path are as well as the effect of this coupling on the performance of the DUT (example: delay of the initial packet of a flow). CPU and Memory Consumption Tests to understand the virtual switch’s footprint on the system, usually conducted as auxiliary measurements with benchmarks above. They include: CPU utilization, Cache utilization and Memory footprint. Future/planned test specs include: Request/Response Performance Tests (TCP, UDP) which measure the transaction rate through the switch. Noisy Neighbour Tests, to understand the effects of resource sharing on the performance of a virtual switch. The flexibility of deployment of a virtual switch within a network means that the BMWG IETF existing literature needs to be used to characterize the performance of a switch in various deployment scenarios. The deployment scenarios under consideration include:

Physical port to virtual switch to physical port

Physical port to virtual switch to VNF to virtual switch to physical port

Physical port to virtual switch to VNF to virtual switch to VNF to virtual switch to physical port

Physical port to virtual switch to VNF

VNF to virtual switch to physical port

VNF to virtual switch to VNF

This section organizes the many existing test specifications into the "3x3" matrix (introduced in ). Because the LTD specification ID names are quite long, this section is organized into lists for each occupied cell of the matrix (not all are occupied, also the matrix has grown to 3x4 to accommodate scale metrics). The tests listed below assess the activation of paths in the data plane, rather than the control plane. (Editor's Note: a complete list of tests is available here: https://wiki.opnfv.org/wiki/vswitchperf_test_spec_review )

Throughput.RFC2889.AddressLearningRate Throughput.RFC2889.AddressCachingCapacity PacketLatency.InitialPacketProcessingLatency

Throughput.RFC2544.SystemRecoveryTime Throughput.RFC2544.ResetTime

Throughput.RFC2889.AddressCachingCapacity

Throughput.RFC2544.PacketLossRate Throughput.RFC2544.PacketLossRateFrameModification Throughput.RFC2544.BackToBackFrames Throughput.RFC2889.ForwardingRate Throughput.RFC2889.ForwardPressure Throughput.RFC2889.BroadcastFrameForwarding RFC2889 Broadcast Frame Latency test

Throughput.RFC2889.ErrorFramesFiltering

Throughput.RFC2544.Soak Throughput.RFC2544.SoakFrameModification

Benchmarking activities as described in this memo are limited to technology characterization of a Device Under Test/System Under Test (DUT/SUT) using controlled stimuli in a laboratory environment, with dedicated address space and the constraints specified in the sections above. The benchmarking network topology will be an independent test setup and MUST NOT be connected to devices that may forward the test traffic into a production network, or misroute traffic to the test management network. Further, benchmarking is performed on a "black-box" basis, relying solely on measurements observable external to the DUT/SUT. Special capabilities SHOULD NOT exist in the DUT/SUT specifically for benchmarking purposes. Any implications for network security arising from the DUT/SUT SHOULD be identical in the lab and in production networks.

No IANA Action is requested at this time.

The authors acknowledge