This shows you the differences between two versions of the page.
Both sides previous revision Previous revision Next revision | Previous revision | ||
wiki:vswitch_work_items [2016/01/06 13:33] Maryam Tahhan [Table] |
wiki:vswitch_work_items [2016/03/03 16:42] (current) Maryam Tahhan |
||
---|---|---|---|
Line 38: | Line 38: | ||
==== Priority 2 ==== | ==== Priority 2 ==== | ||
- | ^ Feature ^ Description ^ Priority ^ JIRA ^ | + | ^ Feature ^ Description ^ Priority ^ JIRA ^ Effort ^ Impact ^ |
- | | Support TCP traffic in traffic gens | Add TCP support to IxNET TCL script. (UDP is hardcoded) | Med | | | + | | Support TCP traffic in traffic gens | Add TCP support to IxNET TCL script. (UDP is hardcoded) | Med | | M | Low | |
- | | TCL script cleanup | Remove TCL lines that can be generated by IxNet generate (highLevelStream 1500+ lines) items (Ongoing) | Med | | | + | | TCL script cleanup | Remove TCL lines that can be generated by IxNet generate (highLevelStream 1500+ lines) items (Ongoing) | Med | | XL | Low | |
- | | Migrate to IXIA python APIs instead of TCL scripts | | Med | | | + | | Migrate to IXIA python APIs instead of TCL scripts | | Med | | XL | High | |
- | | vSwitch None Option | Implement a vswitch None option that allows you to setup the use case you want to test without a vswitch. | Med | | | + | | Detailed benchmark reporting and graphing. | | Med | | XL | High | |
- | | Bypass/SR-IOV setup support | Enable Bypass/SR-IOV setup support with vsperf for the no vswitch use case. | Med | | | + | | IXIA RFC 5481 evaluation and study of scripts | Study the IXIA scripts provided by IXIA and integrate them with the traffic gen utility in VSPERF - scripts are written for the IXIA Sw version supported by POD3 and will need to be tested there. | Med | | L | High | |
- | | Detailed benchmark reporting and graphing. | | Med | | | + | | RFC2889 tests implementation | | Med | | L | Med | |
- | | IXIA RFC 5481 evaluation and study of scripts | Study the IXIA scripts provided by IXIA and integrate them with the traffic gen utility in VSPERF - scripts are written for the IXIA Sw version supported by POD3 and will need to be tested there. | Med | | | + | | deployment as a test argument | remove the deployment from the test name - it's exploding our testcases config file, it should be passed as a parameter/list to the testcase which has an agnostic name and reported as part of the result. | Med | | M | Med | |
- | | RFC2889 tests implementation | | Med | | | + | | Fix OF table structure for PVP and PVVP to match P2P | OF table structure used in VSPERF is not followed by PVP and PVVP scenarios (flows are just in first table) | Med | | XL | Med | |
- | | deployment as a test argument | remove the deployment from the test name - it's exploding our testcases config file, it should be passed as a parameter/list to the testcase which has an agnostic name and reported as part of the result. | Med | | | + | | Allow for conrigurable port numbers for frame modification | physical port numbers are currently hardcoded in frame modification settings, so it can't be used easily with other deployments than P2P | Med | | L | Low | |
- | | Fix OF table structure for PVP and PVVP to match P2P | OF table structure used in VSPERF is not followed by PVP and PVVP scenarios (flows are just in first table) | Med | | | + | | Update implementation for background load generation to run multiple instances of the background load. | background load generated by stress(-ng) doesn't work well with isolcpus option; It has to be modified to run x instances of stress and pinpoint them to "free" cpus. | Med | | L | Med | |
- | | Allow for conrigurable port numbers for frame modification | physical port numbers are currently hardcoded in frame modification settings, so it can't be used easily with other deployments than P2P | Med | | | + | | Xena Networks traffic generator integration with VSPERF | Integrate Xena with VSPERF | Med | | L | High | |
- | | Update implementation for background load generation to run multiple instances of the background load. | background load generated by stress(-ng) doesn't work well with isolcpus option; It has to be modified to run x instances of stress and pinpoint them to "free" cpus. | Med | | | + | | Decouple the controller from OVS | Migrate OpenFlow dependency out of the controllers and down to the vswitches themselves to decouple the controller from OVS | Med | VSPERF-113 | XL | High | |
- | | Xena Networks traffic generator integration with VSPERF | Integrate Xena with VSPERF | Med | | | + | | L2, L3, L4 flows | Extend test definitions and implementation to do L2, L3, and L4 level flow matching | Med | | TENT | High | |
- | | Decouple the controller from OVS | Migrate OpenFlow dependency out of the controllers and down to the vswitches themselves to decouple the controller from OVS | Med | VSPERF-113 | | + | | Test Prioritization | Establish a Wiki/JIRA story where we prioritize the tests to define and implement. | High | | S | High | |
- | | L2, L3, L4 flows | Extend test definitions and implementation to do L2, L3, and L4 level flow matching | Med | | | + | | Common params for traffig gens | leverage the terminology used in RFC 1242 and RFC 2285 to rename the parameters (both input and output) from traffic gens | High | | M | High | |
- | | Test Prioritization | Establish a Wiki/JIRA story where we prioritize the tests to define and implement. | High | | | + | | Support vfio instead of igb_uio | migrate DPDK setup to use vfio or enable vfio as an option | High | | S | High | |
- | | Common params for traffig gens | leverage the terminology used in RFC 1242 and RFC 2285 to rename the parameters (both input and output) from traffic gens | High | | | + | | Report iteration results as well as aggregate results | | High | | S | High | |
+ | |||
==== Priority 3 ==== | ==== Priority 3 ==== | ||
- | ^ Feature ^ Description ^ Priority ^ JIRA ^ | + | ^ Feature ^ Description ^ Priority ^ JIRA ^ Effort ^ Impact ^ |
- | | Single port traffic gen Source and Sink | Single port traffic gen support (today we just use 2 ports) | Low | | | + | | Single port traffic gen Source and Sink | Single port traffic gen support (today we just use 2 ports) | Low | | L | Low | |
- | | pcap replay as an option for traffic gens. | Implement a function to allow for pcap replay from the traffic gens supported by OPNFV (Priority 1 SW pkt Gens, Priority 2 IXIA & Spirent) | Low | | | + | | pcap replay as an option for traffic gens. | Implement a function to allow for pcap replay from the traffic gens supported by OPNFV (Priority 1 SW pkt Gens, Priority 2 IXIA & Spirent) | Low | | L | Med | |
- | | Fix UDP/TCP port error (Port 63 is hardcoded) | | Low | | | + | | Fix UDP/TCP port error (Port 63 is hardcoded) | | Low | | S | Low | |
- | | Linux bridge as a switch | | Low | | | + | | Linux bridge as a switch | | Low | | M | High | |
- | | Kevin's recommendations for perf benchmarking | TODO get these from Kevin | High | | | + | | Kevin's recommendations for perf benchmarking | TODO get these from Kevin | High | | M | Low | |
- | | Ciara's recommendations for vhost benchmarking | TODO get these from Ciara | High | | | + | | Ciara's recommendations for vhost benchmarking | TODO get these from Ciara | High | | M | Low | |
- | | report frameloss % from IXIA | | High | | | + | | Investigate IXIA SDN + NFV testsuite: OpenFlow Switch Forwarding test | Investigate the tests provided by the IXIA blackbook on SDN + NFV testing to see if there are test we can leverge | Low | VSPERF-84 | L | Med | |
- | | qemu support note for vhost-cuse. | | Low | | | + | | RFC 2544 Scalability tech debt - pre intstall flows on the vswitch | | | VSPERF-83 | M | Low | |
- | | Investigate IXIA SDN + NFV testsuite: OpenFlow Switch Forwarding test | Investigate the tests provided by the IXIA blackbook on SDN + NFV testing to see if there are test we can leverge | Low | VSPERF-84 | | + | | Vanilla OVS PVP/PVVP - differences in RX values for bidirectional flows | | Med | VSPERF-117 | L | Med | |
- | | RFC 2544 Scalability tech debt - pre intstall flows on the vswitch | | | VSPERF-83 | | + | | Add Spirent traffic gen virtual equipment to POD3 | | High | VSPERF-143 | | | |
- | | Vanilla OVS PVP/PVVP - differences in RX values for bidirectional flows | | Med | VSPERF-117 | | + | | Add Spirent support for throughput mod vlan family of tests | | Med | VSPERF-145 | | | |
- | | Fully Meshed Tests for vSwitches | needs to be defined | Med | VSPERF-55 | | + | | Add Spirent support for continuous traffic tests | | Med | VSPERF-146 | | | |
- | | Update flow definition and classification fields in LTD | | Low | VSPERF-61 | | + | | Add Spirent support for b2b family of tests | | Med | VSPERF-144 | | | |
- | | Add Spirent traffic gen virtual equipment to POD3 | | High | VSPERF-143 | | + | | Modify Spirent RFC2544 throughput test to use Spirent's REST API | | Med | VSPERF-148 | | | |
- | | Add Spirent support for throughput mod vlan family of tests | | Med | VSPERF-145 | | + | | Add Spirent support for scalability family of tests | | Med | VSPERF-147 | | | |
- | | Add Spirent support for continuous traffic tests | | Med | VSPERF-146 | | + | | Deployment Naming and extension | we need to come up with a better naming scheme for our deployment scenarios, that’s intuitive and extensible, we will work with new committers/contributors on this. | Med | | M | High | |
- | | Add Spirent support for b2b family of tests | | Med | VSPERF-144 | | + | |
- | | Modify Spirent RFC2544 throughput test to use Spirent's REST API | | Med | VSPERF-148 | | + | |
- | | Add Spirent support for scalability family of tests | | Med | VSPERF-147 | | + | |
- | | Test Prioritization | Establish a Wiki/JIRA story where we prioritize the tests to define and implement. | High | | | + | |
- | | Deployment Naming and extension | we need to come up with a better naming scheme for our deployment scenarios, that’s intuitive and extensible, we will work with new committers/contributors on this. | Med | | | + | |
- | | Common params for traffig gens | leverage the terminology used in RFC 1242 and RFC 2285 to rename the parameters (both input and output) from traffic gens | High | | | + | |
==== Activities and Tasks ==== | ==== Activities and Tasks ==== | ||
Line 92: | Line 89: | ||
| | | | | | | | | | ||
| | | | | | | | | | ||
- | |||
- | |||
- | |||