Author
Can networking applications achieve suitable performance with IOMMU at high rates? Our recent PeerJ CS article answers this question by characterizing the performance implications of IOMMU and its cache (IOTLB) on recent Intel Xeon Scalable & AMD EPYC processors at 200 Gbps. Our study shows that enabling IOMMU at high rates could result in an up-to-20-percent throughput drop due to excessive IOTLB misses. Moreover, we present potential mitigation techniques to recover the introduced throughput drop caused by the “IOTLB wall” by using hugepage-backed buffers in the Linux kernel. This is joint work with Alireza Farshin (KTH), Luigi Rizzo (Google), Khaled Elmeleegy (Google), and Dejan Kostic (KTH). Follow the links for PDF and code.”
Author
On April 1, 2023 Professor Maguire became Professor emeritus connected to KTH Royal Institute of Technology, and on April 21 there was a ceremony in Stockholm to mark this event. Congratulations to Professor Maguire on a fantastic career that is nowhere near to being over!
Author
Similarly to multi-core CPUs, also network devices increasingly rely on parallel packet processing engines to achieve insanely high throughput (up to 16 pipes to process 50 terabits per second on a single chip). In our recent paper accepted at ACM SIGMETRICS, we unveil, quantify, and mitigate the impact of deploying existing network monitoring mechanisms on multi-pipe network devices. Our design, called PipeCache, allows to reduce memory requirements (a constrained resource on ASIC devices) up to 16x! A PDF of the paper is available here. Code is available here.
Author
Can one process the equivalent of 1 Tbps of traffic on a single server? In our NSDI’23 paper, we leverage disaggregation principles to push the boundary of what CPU-based packet processors can achieve in terms of throughput for a variety of network functions. For the paper PDF click here. This is a joint work with two visiting doctoral students from Roma tre University. All code is available here.
A video of Tommaso’s NSDI talk:
Author
We are very happy to announce that Alireza Farshin successfully defended his PhD thesis on March 6, 2023! Prof. Gerald Q. Maguire Jr. was a superb co-advisor (as always!), and Prof. Mark Silberstein was a stellar opponent at the defense seminar. Alireza’s thesis is available online:
Realizing Low-Latency Packet Processing on Multi-Hundred-Gigabit-Per-Second Commodity Hardware: Exploit Caching to Improve Performance
This was a hybrid defense, so here is the Zoom screenshot plus an on-site image:
