DASH Workgroup Community Update 1/28/2026

[Kristina Moore]

Hello DASH Open Source Community – thanks for your time yesterday.  Just a reminder that next week is the 1^st Wednesday of the month (which we have agreed to skip). 
I’m also still looking to leverage the Linux Foundation lists more to manage communications.  If you could please take the time to enter your info here, I can proceed with deletion of the sonic-dash@googlegroups list we used when we began the project.

In summary, we reviewed several recently completed and in-progress pull requests, highlighting the work below:

• Progress and timelines for counter standardization, including coordination with AMD on SAI definitions and headers. The group aligned that most or all counters are actively used and should be exposed, with prioritization to be discussed. For the 2026_05 release, SAI headers and related PRs must be merged by end of February ahead of the March code cut, including community discussion and sign‑off. Ownership was clarified across SAI header/libsai development, controller integration, and community coordination.

• Merged the Smart Switch ENI forwarding test PR 21251

• Remove syncd Redis objects when using ZMQ notifications – a critical performance and memory optimization applicable to all DASH configurations PR1694

• HAMGR_D support for rehydration using ProducerStateTable for DASH_HA tables PR563

• Fix DASH meter COUNTERS_DB keys to use VID instead of RID PR1725

• Discussed validation and testing follow‑ups, including the need to validate ASIC DB changes using a 2025_11 build, and ongoing work on switch memory exhaustion tests, kernel panic coverage, and a BFD rewrite on PMON change (approved and awaiting merge).

• Reviewed performance improvements and ZMQ optimizations, with recent merges already delivering significant memory utilization gains on the DPU by removing redundant database caches, and a phased plan outlined for further ZMQ call optimizations.

• Discussed an ICMP redirect / DPU packet handling issue involving unexpected ICMP packets emitted by the DPU. Initial assessment indicates this is not expected behavior under standard configurations. The issue is being taken offline for deeper investigation, with ownership confirmed and additional stakeholders looped in. Potential mitigations and additional test support options were noted.

For Complete Details, please see the “Full DASH Community Notes” near the end of this communication. 

 

---

In Summary (full list below), since the last Community call we have:

19 PRs Completed (+14)

12 in To Do (+1 )

8 in Draft (+2)

35 in Progress (-10)

11 Awaiting Review (-2)

6 Awaiting Merge (+/-0) *new category

 

Just a reminder that we would encourage/invite Community members to present to the Community (test runs or progress, new scenarios, etc…), just ‘r’ to let me know, or generate a PR in the repo.

The DASH channel link is here to subscribe / access WG content (and click the bell to receive notifications). 

 Thank you for your time/contributions – see you on 2/11/2026

 

Meeting Title:  SONiC-DASH-Workgroup Community Meeting #169

Attendees (13):

DASH Group to join: https://groups.google.com/g/sonic-dash

Linux Foundation list: https://lists.sonicfoundation.dev/g/SONiC-Dash

 

Don Ewald - Cisco	Michael Aronovici - Cisco	Prince Sunny - MSFT	Vivek Reddy Karri - Nvidia
Abdul Rouff Basheer - Nvidia	Michal Zygmunt - MSFT	Ramesh Raghupathy - Cisco
Gagan Punathil Ellath - Nvidia	Mircea Dan Gheorghe - Keysight	Swami Balasubramanian - Keysight
Kristina Moore - MSFT	Murali Venkateshaiah - Cisco	Veerappan, Senthilnathan - AMD

   

Full DASH Community Notes 😊

•  Counter Standardization Progress and Timelines: discussed the status, priorities, and expected timelines for the counter standardization effort, including coordination with the AMD team and the need for community sign-off before the March code cut.

• Project Status and Team Coordination: Murali explained that the team has discussed with the AMD group and plans to start work on SAI definitions, SAI header definitions, and related tasks. Senthil confirmed the plan, and Murali noted that the work has not formally started but will begin soon, with Vijay and Senthil preparing resources.
• Exposure and Prioritization of Counters: Michal clarified that the goal is to expose the counters that AMD has, noting that prioritization and scoping will be required. Kristina asked Michal if all counters are needed immediately or if prioritization is possible; Michal responded that most, if not all, are actively used and needed.
• Timeline and Community Process: Michal inquired about high-level timelines, and Murali stated that the timeline is not yet finalized. Senthil indicated that more details would be available by the next call. Prince provided additional context, stating that for the 2026_05 release, the code cut is around March, so SAI headers should be defined and PRs merged by the end of February, including community discussions and sign-off.
• Roles and Next Steps: Murali summarized that Vijay would develop the SAI header and libsai functionality, while Michal and others would handle pulling counters into their controller. Prince agreed to coordinate with the community and help as needed.

 

• Recent Pull Requests and Feature Updates: reviewed several recently completed and in-progress pull requests, highlighting work on Smart Switch ENI forwarding, memory optimizations, HA features, and bug fixes contributed by various team members and organizations.

• Smart Switch ENI Forwarding Test: noted that the Smart Switch ENI based forwarding test (PR 21251) was merged.
• Memory Optimization and ZMQ Notifications: PR1694 to remove syncd Redis objects when using ZMQ notifications was discussed. This is a generic improvement for all DASH configurations, improving performance and memory usage. Prince emphasized that this PR is critical and a significant improvement.
• Validation and Testing of ASIC DB Changes: Murali asked if the ASIC DB changes were validated with tests. Prince explained that a build with version 2025_11 is needed for validation, and the team is working on creating this build. Murali and Senthil agreed to include this validation in their to-do list.
• HA and Table Logic Enhancements: PR563 highlighted work on HAMGR_D to support rehydration using ProducerStateTable for DASH_HA_tables, with contributions from Jing and the Keysight team.
• PR1725 for correcting DASH meter COUNTERS_DB keys to use VID instead of RID, with Prince and Mukesh involved.
• Other Notable PRs and Tests: Other items mentioned included switch memory exhaustion and kernel panic tests, and a new PR for BFD rewrite on PMON change, which was approved and awaiting merge. Also mentioned ongoing work on HA Private Link support and DPU FDB cleanup for restart scenarios.

 

• Performance Improvements and ZMQ Optimization: discussed recent and ongoing performance improvements related to ZMQ calls, memory utilization, and redundant cache removal, with Prince outlining a phased approach and highlighting significant gains from recent merges.

• ZMQ Call Performance Issues: Michal asked about performance problems with ZMQ calls. Prince responded that improvements have been made recently, but further optimizations are still in the pipeline.
• Memory Utilization Gains: Prince explained that recent merges have significantly improved memory utilization by removing redundant caches in the database on the DPU, leading to a substantial reduction in memory usage.

 

• ICMP Redirect and DPU Packet Handling Investigation: Mircea raised a technical question about unexpected ICMP packets sent by the DPU, leading to a detailed discussion to clarify expected behavior, diagnose the issue, and coordinate follow-up investigation.

·        Issue Description and Initial Analysis: Mircea described a scenario where the DPU sends ICMP packets wrapping VXLAN and TCP payloads, which was not expected. Michal and Prince clarified that there is no standard use case for the DPU to send such ICMP packets unless related to a custom HA protocol.

·        Technical Details and Packet Inspection: Mircea provided details about the packet structure and the DPU image version. Abdul and Prince discussed possible causes, including missing table programming or private link transformations, and compared the behavior to previous image versions.

·        Next Steps and Ownership: confirmed that the issue is under discussion via email to Abdul, with Mircea, Michael, and Prince included. Michal suggested involving Maria.  Abdul agreed to take the investigation internal.

·        Potential Workarounds and Additional Support: Senthil mentioned that disabling the ENI object field for ICMP direct errors could be an option if supported by Nvidia, and offered to provide a new GNMI container image for further testing.

·        OCP Europe Event Discussion: briefly discussed the OCP Europe event, with consensus that attending would be interesting but challenging for those based on the West Coast

 

 

 

Sticky for Links/Reference:
DASH Groups to join to receive Invites, Meeting Notes, and Comms	DASH: https://groups.google.com/g/sonic-dash     DASH-Test-Workgroup Group: https://groups.google.com/g/sonic-dash-test-workgroup   Linux Foundation list: https://lists.sonicfoundation.dev/g/SONiC-Dash	If anyone knows potentially interested people who would like info re: our community, please have them joins these groups for receive Comms, etc…
Recordings	Teams: SONiC-DASH Workgroup Community Meeting-20260128_090402-Meeting Recording.mp4 DASH Community https://youtu.be/bY55C44HN-o	1/28/2026 DASH Community Call; please request access via the link if you are not able to view/listen
Azure DASH GitHub Repo:	https://github.com/sonic-net/DASH
Test/Docs folder:	https://github.com/sonic-net/DASH/blob/main/test/docs/dash-test-workflow-saithrift.md	Ideal test workflow is here, converted to .md
SAI Thrift	SAI Thrift PR	Client server needed for testing
P4	https://opennetworking.org/p4/ and https://p4.org/working-groups/	Open source, domain-specific programming language for network devices, specifying packet processing for data plane devices (switches, routers, NICs, filters, etc.)
PINS	https://opennetworking.org/pins/
PNA consortium spec	https://p4.org/p4-spec/docs/PNA-v0.5.0.html	An architecture describing the structure and common capabilities of network interface controller (NIC) devices which process packets transiting one or more interfaces and a host system. Describes the structure and capabilities of the pipeline, and a user program, which specifies the functionality of the programmable blocks within that pipeline. For more information, see the P4 Language Consortium specifications
IPDK	Infrastructure Programmer Development Kit (ipdk.io) and https://github.com/ipdk-io/ipdk-io.github.io	IPDK is an open source, vendor agnostic framework of drivers and APIs for infrastructure offload and management which runs on a CPU, IPU, DPU or switch. IPDK runs in Linux and uses a set of well-established tools such as DPDK and P4 to enable network virtualization.
bmv2	https://github.com/p4lang/behavioral-model	The second version of the reference P4 software switch, nicknamed bmv2 (for behavioral model version 2). The software switch is written in C++11. It takes as input a JSON file generated from your P4 program by a P4 compiler and interprets it to implement the packet-processing behavior specified by that P4 program
DPDK	https://www.dpdk.org/	DPDK is the Data Plane Development Kit which consists of libraries to accelerate packet processing workloads running on a wide variety of CPU architectures.
Linux Foundation SmartSwitch	https://lists.sonicfoundation.dev/g/sonic-smartswitch/calendar

  

Thank you again for your participation…

Kristina Moore MBA, M.S., CISSP - Azure Core Principal PM / DASH & SmartSwitch
Office: 425-722-7720     Mobile: 425-876-2040     Email: kri...@microsoft.com
DASH Group to join: https://groups.google.com/g/sonic-dash    
Linux Foundation:  https://lists.sonicfoundation.dev/g/SONiC-Dash