Udhaya chandran Shanmugam
Dell Technologies
Actions
Technical leader in both traditional and open network architectures with 16+ years of experience working in various network segments like Datacenter, Enterprise, ISP and cloud networking.
Area of Expertise
Topics
Optimizing Data Center Scalability: Implementing EVPN VxLAN Multisite for GPUaaS AI Clusters
Scaling data centers for GPU-as-a-Service (GPUaaS) customers- connecting multiple data halls poses significant challenges- especially as the number of VxLAN endpoints grows within the network fabric. This presentation explores how EVPN VxLAN Multisite can address these challenges by dividing the network fabric into smaller segments and isolating them with split-horizon groups. We'll discuss the standard implementation and how SONiC- combined with Free Range Routing (FRR) and Switch Abstraction Interface (SAI)- achieves this. Additionally- we'll cover important considerations for network design to avoid common pitfalls and ensure a robust- scalable infrastructure for AI clusters.
Network impact for Distributed Inferencing
Distributed inferencing performance is critical for the broader adoption of AI applications in enterprises. To support this, SONiC must be equipped to handle the demands of AI scale-out environments. This talk outlines the key performance indicators (KPIs) necessary for SONiC readiness—specifically latency, throughput, and reliability. It also explores the role of disaggregated inferencing in current and future AI deployments, and how SONiC can be effectively utilized to support both prefill and decoding networks in distributed inference architectures.
Improving AI Workload Traffic Management with Adaptive Routing in SONiC
The rapid growth of AI workloads in data centers and cloud environments has introduced new challenges in network traffic management, particularly in achieving efficient load balancing across high-bandwidth links. Traditional static routing mechanisms often fall short in handling the bursty and uneven nature of AI traffic, leading to congestion, increased latency, and underutilized resources.
Enhancing Storage AI Fabrics: Comparing EVPN MultiHoming and MC-LAG for Improved Scalability & Mgmt.
In this presentation, we explore how Ethernet VPN (EVPN) MultiHoming offers a compelling alternative to Multi-Chassis Link Aggregation (MC-LAG) in the context of expanding storage and access networks within the AI Fabric. With the existing MC-LAG MultiHoming solution in SONiC, introducing the standards-based EVPN MultiHoming provides another robust option. This session delves into the critical comparison between these two solutions, examining their scalability, convergence, and management. Through detailed analysis, we highlight the advantages and drawbacks of each approach and the best choice for the specific needs. Additionally, we discuss practical use cases and effective troubleshooting techniques, ensuring a comprehensive understanding of how EVPN MultiHoming can reduce operational overhead and enhance network efficiency.
Udhaya chandran Shanmugam
Dell Technologies
Actions
Please note that Sessionize is not responsible for the accuracy or validity of the data provided by speakers. If you suspect this profile to be fake or spam, please let us know.
Jump to top