8 papers
Ofan: Optimal Load Balancing for AI Training
Sarah McClure, Evyatar Cohen, Jakob Krebs +5
The extreme collective completion time (CCT) demands of AI workloads challenge existing packet spraying algorithms, which can have trouble efficiently load-balancing workloads that…
Incast-Free MoE Rate-Based Scheduling
Evyatar Cohen, Jose Yallouz, Alexander Shpiner +3
Mixture of Experts (MoE) architectures have become key to large language models; however, their typical round-robin (RR) scheduling introduces significant bottlenecks. In this pape…
CAPS: Fine-Tuning CCA Timing
Raphael Zailer, Isaac Keslassy
Data-center congestion control targets high throughput, fair bandwidth allocation, and low latency. Modern transports couple rate computation and packet scheduling into a single fe…
Understanding Cross-Cloud Interconnects: Hands-On Measurements and Cost Optimization
Eitan Eliav, Isaac Keslassy, David Breitgand +2
New services such as Google Cross-Cloud Interconnect (CCI) address the rise in fast and large-scale cross-cloud data transfers. CCI offers dedicated high-throughput links with low…
Scheduling Parallel Optical Circuit Switches for AI Training
Kevin Liang, Litao Qiao, Isaac Keslassy +1
The rapid growth of AI training has dramatically increased datacenter traffic demand and energy consumption, which has motivated renewed interest in optical circuit switches (OCSes…
Scaling Routers with In-Package Optics and High-Bandwidth Memories
Isaac Keslassy, Ilay Yavlovich, Jose Yallouz +3
This paper aims to apply two major scaling transformations from the computing packaging industry to internet routers: the heterogeneous integration of high-bandwidth memories (HBMs…
Congestion Control for Spraying with Congested Paths
Barak Gerstein, Mark Silberstein, Isaac Keslassy
Packet spraying approaches are increasingly deployed in datacenter networks. However, their combination with existing congestion control algorithms (CCAs) may lead to poor QoS, esp…
2SYN: Congestion-Aware Multihoming
Kfir Toledo, Isaac Keslassy
When sending flows to arbitrary destinations, current multihoming routers adopt simple congestion-oblivious mechanisms. Therefore, they cannot avoid congested paths. In this paper,…