June 2024 - Submissions-0x18

Re: [Netdev 0x18] Submission #7 "Characterizing IOTLB Wall for..."
by Alireza Farshin 04 Jul '24

04 Jul '24

Hi, Thank you so much for the information. I can record the video by myself. When is the deadline for submitting it? And would it be okay to provide a Dropbox link in the hotcrp? Best Regards, Alireza From: submissions-0x18(a)netdevconf.info <submissions-0x18(a)netdevconf.info> Date: Monday, 24 June 2024 at 14:12 To: Alireza Farshin <afarshin(a)nvidia.com> Cc: jhs(a)mojatatu.com <jhs(a)mojatatu.com>, lael.santos(a)expertisesolutions.com.br <lael.santos(a)expertisesolutions.com.br> Subject: [Netdev 0x18] Submission #7 "Characterizing IOTLB Wall for..." External email: Use caution opening links or attachments Dear Alireza Farshin, We are contacting you because your netdev 0x18 session Characterizing IOTLB Wall for Multi-100-Gbps Linux-based Networking is marked as being presented remotely. For remote speakers we ask to have the presentation prerecorded as to avoid technical issues during the event. We provide some time slots to help in this task in here: https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fcalendar.…<https://calendar.app.google/AqbKvEhqeiFTbAdj7> If you are willing to record the presentation video by yourself, please be sure to checklist the following criteria: - Minimum resolution should be Full HD (1080p) in widescreen format. - Please be sure to capture good audio, this is even more important than video to have your message reach the audiance. - Please record yourself in a clean background and with enough light as to be easily distinguishable. Please be sure to not cover important content from your slides. - Any codec and video container is fine. People tend to like h264 for video and aac for audio. - If you need references, please look at netdev 0x17 remote sessions: - XDP Offload using Nanotube - by Neil Turton (https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fyoutu.be%…<https://youtu.be/e8zv-F2VL5g?t=113>) - Is scaling eBPF easy yet? - by Ivan Koveshnikov (https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fyoutu.be%…<https://youtu.be/TGgIT9rMsns>) - Synce4l: open-source implementation of Synchronous Ethernet - by Michał Michalik and Arkadiusz Kubalewski (https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fyoutu.be%…<https://youtu.be/sbR-pf3sWuc>) Contact Netdev Society <submissions-0x18(a)netdevconf.info> with any questions or concerns. - Netdev 0x18 Submissions

2 3

Re: [Netdev 0x18] Accepted submission #37 "Interrupt moderation approaches for vdpa..."
by Parav Pandit 15 Jun '24

15 Jun '24

Hi, Unfortunately, I will not be able to present on this topic due to unforeseen circumstances. I kindly request to withdraw this submission. Regards, Parav Pandit > -----Original Message----- > From: submissions-0x18(a)netdevconf.info <submissions- > 0x18(a)netdevconf.info> > Sent: Tuesday, June 4, 2024 3:38 AM > To: Parav Pandit <parav(a)nvidia.com> > Cc: submissions-0x18(a)netdevconf.info > Subject: [Netdev 0x18] Accepted submission #37 "Interrupt moderation > approaches for vdpa..." > > Dear Parav, > > The Netdev 0x18 program committee is delighted to inform you that your > submission #37 has been accepted to appear in the conference. > > Title: Interrupt moderation approaches for vdpa device > Authors: Parav Pandit > Dragos Tatulea > Site: > https://0.0.0.24/. > netdevconf.info%2Fpaper%2F37%3Fcap%3D037a5MxDoHUisA4&data=05%7 > C02%7Cparav%40nvidia.com%7C236ca323150d4500012a08dc841991e8%7C4 > 3083d15727340c1b7db39efd9ccc17a%7C0%7C0%7C638530492563548164%7 > CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJB > TiI6Ik1haWwiLCJXVCI6Mn0%3D%7C0%7C%7C%7C&sdata=NTAxD61RN23lv2g > WlBaPmHxoGX6b%2BOlFTkjEC4dOOrE%3D&reserved=0 > > Congratulations! > > Please login and look at the comments left behind for you. > > Contact Netdev Society <submissions-0x18(a)netdevconf.info> with any > questions or concerns. > > - Netdev 0x18 Submissions

1 0

Re: [Netdev 0x18] Accepted submission #11 "vDPA Live Migration Downtime..."
by Dragos Tatulea 14 Jun '24

14 Jun '24

Hi, Thank you for reaching back. Please see below for more. On Fri, 2024-06-14 at 16:37 +0000, submissions-0x18(a)netdevconf.info wrote: > Dear Dragos, > > In your submission: > Title: vDPA Live Migration Downtime improvements for net devices > Authors: Eugenio Perez Martin (RedHat) > Dragos Tatulea (NVIDIA) > Si-Wei Liu (Oracle) > Site: https://0x18.netdevconf.info/paper/11?cap=011ars5NOJJDNbY > > You mention that you are looking to have a round table discussion. Should > we make this session a BoF and give you one hour instead of a talk which > will allocate 30 minutes. > We fear that 1 hour might be too long. If we could do a 30 minute talk + 10-15 minutes of questions/discussion it would be great. Would that be an option? Thanks, Dragos

1 0

Re: [EXTERNAL] [Netdev 0x18] Abstract for #1 "A Linux Implementation of SEARCH -- a..."
by Li, Feng 14 Jun '24

14 Jun '24

Thanks for the heads up. I updated the abstract and just in case, I put the plaintext version here. TCP slow start is designed to ramp up to the network congestion point quickly, doubling the congestion window each round-trip time until link capacity is reached, whereupon TCP exits the slow start phase. Unfortunately, the default Linux TCP slow start implementation -- TCP Cubic with Hystart [1] -- can cause premature exit from slow start, especially over wireless links, degrading link utilization. However, without HyStart, TCP exits too late, causing unnecessary packet loss. To improve TCP performance during slow start, we propose the Slow start Exit At Right CHokepoint (SEARCH) algorithm [2] where the congestion point is determined based on bytes delivered compared to the expected bytes delivered, smoothed to account for link latency variation and normalized to accommodate link capacities. We implemented SEARCH as a Linux kernel v5.16 module\footnote{We will publish source code for pubic review and evaluation} and evaluated it over WiFi, 4G/LTE, and low earth orbit (LEO) and geosynchronous (GEO) satellite links. Analysis of the results show that the SEARCH reliably exits from slow after the congestion point is reached but before inducing packet loss. [1] Ha, S., and Rhee, I. 2011. Taming the elephants: New tcp slow start. Computer Networks 55(9):2092–2110. [2] Kachooei, M. A.; Chung, J.; Li, F.; Peters, B.; and Claypool, M. 2023a. Search: Robust tcp slow start performance over satellite networks. In 2023 IEEE 48th Conference on Local Computer Networks (LCN), 1–4. -- On 6/14/24, 6:34 AM, "submissions-0x18(a)netdevconf.info <mailto:submissions-0x18@netdevconf.info>" <submissions-0x18(a)netdevconf.info <mailto:submissions-0x18@netdevconf.info>> wrote: Dear Feng, Could you please fixup up the abstract references. They dont show up well when rendered on a web page. Contact Netdev Society <submissions-0x18(a)netdevconf.info <mailto:submissions-0x18@netdevconf.info>> with any questions or concerns. - Netdev 0x18 Submissions

1 0

[Netdev 0x18] Accepted submission #11 "vDPA Live Migration Downtime..."
by submissions-0x18＠netdevconf.info 14 Jun '24

14 Jun '24

Dear Si-Wei, In your submission: Title: vDPA Live Migration Downtime improvements for net devices Authors: Eugenio Perez Martin (RedHat) Dragos Tatulea (NVIDIA) Si-Wei Liu (Oracle) Site: https://0x18.netdevconf.info/paper/11?cap=011ars5NOJJDNbY You mention that you are looking to have a round table discussion. Should we make this session a BoF and give you one hour instead of a talk which will allocate 30 minutes. Contact Netdev Society <submissions-0x18(a)netdevconf.info> with any questions or concerns. - Netdev 0x18 Submissions

1 0

[Netdev 0x18] Accepted submission #11 "vDPA Live Migration Downtime..."
by submissions-0x18＠netdevconf.info 14 Jun '24

14 Jun '24

Dear Eugenio, In your submission: Title: vDPA Live Migration Downtime improvements for net devices Authors: Eugenio Perez Martin (RedHat) Dragos Tatulea (NVIDIA) Si-Wei Liu (Oracle) Site: https://0x18.netdevconf.info/paper/11?cap=011ars5NOJJDNbY You mention that you are looking to have a round table discussion. Should we make this session a BoF and give you one hour instead of a talk which will allocate 30 minutes. Contact Netdev Society <submissions-0x18(a)netdevconf.info> with any questions or concerns. - Netdev 0x18 Submissions

1 0

[Netdev 0x18] Accepted submission #11 "vDPA Live Migration Downtime..."
by submissions-0x18＠netdevconf.info 14 Jun '24

14 Jun '24

Dear Dragos, In your submission: Title: vDPA Live Migration Downtime improvements for net devices Authors: Eugenio Perez Martin (RedHat) Dragos Tatulea (NVIDIA) Si-Wei Liu (Oracle) Site: https://0x18.netdevconf.info/paper/11?cap=011ars5NOJJDNbY You mention that you are looking to have a round table discussion. Should we make this session a BoF and give you one hour instead of a talk which will allocate 30 minutes. Contact Netdev Society <submissions-0x18(a)netdevconf.info> with any questions or concerns. - Netdev 0x18 Submissions

1 0

[Netdev 0x18] Abstract for #1 "A Linux Implementation of SEARCH -- a..."
by submissions-0x18＠netdevconf.info 14 Jun '24

14 Jun '24

Dear Maryam Ataei, Could you please fixup up the abstract references. They dont show up well when rendered on a web page. Contact Netdev Society <submissions-0x18(a)netdevconf.info> with any questions or concerns. - Netdev 0x18 Submissions

1 0

[Netdev 0x18] Abstract for #1 "A Linux Implementation of SEARCH -- a..."
by submissions-0x18＠netdevconf.info 14 Jun '24

14 Jun '24

Dear Feng, Could you please fixup up the abstract references. They dont show up well when rendered on a web page. Contact Netdev Society <submissions-0x18(a)netdevconf.info> with any questions or concerns. - Netdev 0x18 Submissions

1 0

Re: [Netdev 0x18] Accepted submission #41 "Shared Memory Pool for Representors"
by William Tu 13 Jun '24

13 Jun '24

Hi Program Committee members, In response to feedback “very Nvidia specific”, I talked to Intel and they are willing to collaborate on the efforts. I want to see if I can change about Authors and add “Michal Swiatkowski” from Intel. (I will also update at the submission page) We will present from both the ICE driver and MLX5 driver’s point of view. Thank you William From: submissions-0x18(a)netdevconf.info <submissions-0x18(a)netdevconf.info> Date: Monday, June 3, 2024 at 3:14 PM To: William Tu <witu(a)nvidia.com> Cc: submissions-0x18(a)netdevconf.info <submissions-0x18(a)netdevconf.info> Subject: [Netdev 0x18] Accepted submission #41 "Shared Memory Pool for Representors" External email: Use caution opening links or attachments Dear William, The Netdev 0x18 program committee is delighted to inform you that your submission #41 has been accepted to appear in the conference. Title: Shared Memory Pool for Representors Authors: William Tu (nvidia) Yossi Kuperman Site: https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2F0x18.netd…<https://0x18.netdevconf.info/paper/41?cap=041aNG6hM6H92qU> Congratulations! Please login and see further comments. The main concern from the PC seems to be that this is very vendor specific an may not be relatable to other vendor devices. In addition there seems to be some vendor specific usage of words such as "channel" that may need some disambiguation. Suggestions of the PC are to spend some more time explaining the proposed solution in the abstract. Also some clarity on "channel" might be useful as it wasn't clear if that is a queue grouping or a switchdev abstract. Also is there any drivers other than this one that would benefit from these changes, the concern of the committee is that this is very Nvidia specific and will likely receive strong pushback. Contact Netdev Society <submissions-0x18(a)netdevconf.info> with any questions or concerns. - Netdev 0x18 Submissions

1 1