Researcher at NVIDIA. Interested in computer systems, cinema, and music.
Joined March 2019
- Tweets 28
- Following 127
- Followers 90
- Likes 1,673
Photos and videos
Alireza Farshin retweeted
11 Dec 2024
1/3
At CoNEXT ’24, @Skazza_94 presented our FAJITA paper for running stateful network functions at very high rates (~1.4 Tbps). PDF video are at the @ERC_Research CoG ULTRA blog post below. Click below to see something perhaps unexpected in this work.
bit.ly/4g63His
1
3
9
482
Alireza Farshin retweeted
31 Jul 2024
New on #PACMNET "NetConfEval: Can LLMs Facilitate Network Configuration?" investigates opportunities and challenges in operating network systems using LLMs. #LLM #NetConfEval @Skazza_94 @dmkostic @marchiesa @wangchangjie13 @si_ferlin @alirezafarshin bit.ly/3VZtHDn
5
8
510
Alireza Farshin retweeted
24 Jul 2024
We show that FAJITA can process up to 170Mpps with an NF consisting of a flow counter a rate limiter and a load balancer. We then show how to improve by 50% the throughput of NF chains that match on different fields of the header, processing up to 90Mpps.
1
1
4
332
Alireza Farshin retweeted
24 Jul 2024
We address these challenges by devising a careful NF packet processor (on top of FastClick), which we call FAJITA, that carefully orchestrate the prefetching and disaggregation of any information that must be retrieved by any core to process a packet.
1
1
4
181
Alireza Farshin retweeted
24 Jul 2024
We discovered that:
- Chasing a single pointer on a CPU may cost you 10 Mpps
- Many NF chains retrieve flow states based on different packet header fields, breaking the "exclusive memory access" guarantees of high-speed multi-core CPUs
In our CoNEXT'24 paper... (see comments)
1
2
19
1,016
Alireza Farshin retweeted
18 Jul 2024
Alireza Farshin and Luigi Rizzo on "Characterizing IOTLB Wall for Multi-100-Gbps Linux-based Networking" at #netdevconf 0x18 @netdev01
1
12
353
Alireza Farshin retweeted
8 May 2024
Do large language models facilitate network configuration? We investigate opportunities and challenges in operating network systems using LLMs in our accepted ACM CoNEXT'24 paper. Thanks, @WangChangjie13 @Skazza_94 @alirezafarshin @si_ferlin @dmkostic. More info in the comments.
1
4
19
1,090
Alireza Farshin retweeted
18 Nov 2023
KTH held a superb ceremony at the Stockholm City Hall yesterday, and @alirezafarshin received his PhD diploma!
2
2
23
1,084
Alireza Farshin retweeted
28 Jun 2022
Our SSF Time-Critical Clouds project is ending and this article describes our efforts on reducing latency for high speed networking services (in Swedish) Hint: it's all about the cache! framtidensforskning.se/2022/… @Amir_Roozbeh @alirezafarshin @GQMaguireJr #research #ssf #latency
3
6
Alireza Farshin retweeted
14 Jun 2022
I have been so fortunate to work with superb doctoral students and postdocs, but of course they graduate and get new jobs. The updated job openings for multiple doctoral student and postdoc positions in my ERC ULTRA project are below:
kth.se/blogs/ultra/2020/11/n…
8
12
Alireza Farshin retweeted
4 Apr 2022
Hugely honored that our "Packet Order Matters" paper received the Community Award at NSDI 2022! #ercfunded #eufunded #nsdi22 #usenix @ERC_Research @TomBarbette @hamidgh09 @GQMaguireJr @marchiesa @alirezafarshin @girondi_massimo @Amir_Roozbeh usenix.org/conference/nsdi22…
1
7
56
Alireza Farshin retweeted
25 Mar 2022
Check out our recently published paper (I missed the notification :p) in IEEE/ACM ToN where we present a technique to combine state among multiple NFs and surgically modify flows at different layers at high speed (already flashed at 100G :) ) diva-portal.org/smash/get/di…
1
3
Alireza Farshin retweeted
29 Jan 2022
Our upcoming NSDI 2022 paper Packet Order Matters shows a surprising result: deliberately delaying packets can improve the performance of backend servers by up to about a factor of 2! #ercfunded #eufunded
people.kth.se/~dejanko/docum…
kth.se/blogs/ultra/2021/11/o…
6
14
Alireza Farshin retweeted
21 Jan 2022
We will present Morpheus at ACM ASPLOS! We show that it is possible to optimize eBPF/XDP packet-processing programs at run-time.
Here, an abstract that recap our idea: sebymiano.github.io/publicat…
We also received the 3 ACM badges for reproducibility:
github.com/Morpheus-compiler…
1
13
56
18 Apr 2021
If you want to know more about our @ASPLOSConf paper about optimizing packet processing frameworks, join our session tomorrow (April 19th) at 16:30 CEST. Check packetmill.io for the paper and the video with English/Farsi/French subtitles. #PacketMill @TomBarbette
1
1
18
Alireza Farshin retweeted
14 Apr 2021
I would appreciate your help in spreading the word about the extended deadline (May 13) for our Assoc. Prof. in Systems job opening:
kth.se/en/om/work-at-kth/led…
Please do not hesitate to contact Dejan Kostic if you have any questions.
17
14
Alireza Farshin retweeted
24 Mar 2021
Do you want to know more about SmartNIC (CPU-based) performance?
Come to see what Georgios and
@TomBarbette
discovered at their PAM talk on Mar 30, 7pm CEST!
One spoiler: multi-table packet processing and rule insertions are expensive!
Joint work with D. Kostic and G. Maguire
23 Mar 2021
#PAM2021 will start next Monday, March 29. An exiting program with 2 keynotes by @cristel_pelsser and @kgummadi at free registration is awaiting you:
pam2021.b-tu.de/program/
Register for free!
1
1
9
Alireza Farshin retweeted
16 Feb 2021
The video (video.fosdem.org/2021/D.sdn/…) of our last talk at FOSDEM'21 is available! We present FastClick and discuss the future of sw dataplanes (spoiling a bit PacketMill, to be presented at ASPLOS) ➡️tombarbette.be/fosdem21 #fastclick #fosdem21 #packetmill @alirezafarshin
2
1
15 Jul 2020
If you want to learn more about Direct Cache Access (DCA) or Intel Data Direct I/O (DDIO) technology when shifting toward multi-hundred-gigabit networks, check out our @usenix ATC'20 paper. #ATC20 #USENIX
usenix.org/conference/atc20/…
8
13