Data centers workloads are evolving rapidly – driven by high volumes of end users, application types, cluster nodes, and overall data movement in the cloud.
In turn, the cross-sectional bandwidth of these cloud-scale data center networks is quickly increasing, outpacing the increase in physical link speeds. Traditionally, both ECMP (Equal Cost Multipathing) and port-channel implementations attempt to distribute flows uniformly across physical links that form the logical path. Deciding which flows use which physical link has been traditionally based on a static hash of a fixed set of fields from the packet header. This static hashing scheme is not ideal for high and diverse cloud traffic patterns, causing network polarization whereby multiple traffic flows may traverse and burden the same link while leaving other links underutilized. A range of flexible hashing enhancements has emerged to provide an alternative, improving network performance and overcoming limitations imposed by traditional hashing schemes.
Data Center Traffic Drives a New Look at Hashing Mechanisms
Traditional load balancing systems split traffic bound through a logical fat link to multiple outgoing physical links. Typically, the physical link corresponding to a flow is ascertained by calculating a hash based on packet header fields and a subsequent modulo operation based on the number of physical links. A good load balancing system should be able to evenly split the traffic to the multiple outgoing links. In addition, packets belonging to the same flow should flow out in order to the end destination.
This static hashing scheme has worked well in carrier and enterprise networks, but is proving ineffective for data centers and cloud networks. Web, application, and database server applications running as VMs (virtual machines) that can reside in any server in any rack – coupled with the growing use of clustered applications (such as Hadoop) in modern data centers – has increased east-west traffic patterns in data center networks. East-west traffic includes server-to-server, server-to-storage, and server rack-to-server rack. This is fundamentally changing the design of network topologies, from oversubscribed and tiered networks to fast, fat, and flat networks which demand new features in network switches.
Driven by the latest silicon advances, increased bandwidth and port densities of switch-on-a-chip systems, traditional 3-tier network designs are quickly being replaced by these fast, fat and flat networks – comprised of resilient and flexible CLOS topologies with very high cross-sectional bandwidth.
For designers, this transition has some important factors to consider. Data centers of massive scale, such as those using several thousand links, will experience frequent link failures resulting in network polarization. Yet deployed networks must perform normally and deliver packets in order, even under such failure conditions. Further, as newer protocols and encapsulations are introduced as a means to improve data center automation and network management, newer packet header fields redefine flows and how packets need to be treated. Along with the introduction of new network features, network operators must have the general ability to debug and trace packets. Lastly, with the increased adoption of cloud hosting services, security is more prominent and essential than ever in network operations. Stateful packet inspection and intrusion detection systems will certainly continue to gain importance.
Fueled by these requirements, network operators are turning to switch architectures optimized for cloud networking such as the Broadcom StrataXGS architecture that features Smart-Hash technology, , engineered specifically to addresses the intelligent hashing needs of high-performance cloud network designs today and moving forward. Smart-Hash enables Resilient Hashing to address link failure and live topology changes; Flexible Hashing to handle newer encapsulations and protocols; and Symmetric Hashing to incorporate packet traceability and stateful protocol debug.
Intelligent Hashing Improves Cloud Network Visibility and Performance
Data center networks are driving innovation, fueled by the unprecedented growth of recent years. Large enterprises are building enormous data centers with massive scale. Others are choosing to host their data centers in the cloud. But it’s the smarter and more sophisticated hashing technologies that provide greater flexibility and network visibility – without these advances, data centers of today’s massive scale may suffer from inefficiencies and performance challenges when deploying new and rapidly evolving protocols.
Traditional static hashing schemes work well for enterprise and carrier networks, yet the same approach – often adopted for data center networks – is proving the need for smarter solutions such as Broadcom’s Smart-Hash technology. More intelligent solutions, optimized for cloud network environments and performance demands, are essential to managing the requirements imposed by current trends in cloud and data center networking. This type of advanced technology offers an alternative to design approaches based on static hashing schemes, and enables ideal application performance under typical data center traffic loads. Cloud network operators already face daunting challenges in scaling their network infrastructure to tomorrow’s workloads – alternative hashing options support the need for cloud network flexibility, now and in the future.
About the Authors
Sujal Das serves as Director of Product Marketing for Broadcom Corporation’s Infrastructure and Networking Group (ING). In this role, Das is responsible for driving Broadcom’s Ethernet switch business in the data center and enterprise LAN market segments, and the development of product, ecosystem and strategy based on technology trends and application workloads. Das earned a Bachelor of Science EEE degree from the Birla Institute of Technology and Science in Pilani, India and an MBA from Santa Clara University. He has published and presented on numerous data center networking and virtualization related topics in well-known industry publications and events. Das has been active in the open source community, driving multiple initiatives that have helped proliferate the adoption of high-performance networking in data center applications.
Rochan Sankar serves as Associate Director of Product Line Management for Broadcom Corporation’s Infrastructure and Networking Group (ING). Sankar has 13 years of experience in defining and managing leading-edge semiconductor products for the networking and communications industries. Prior to Broadcom, he held senior positions in chip architecture, strategic marketing, and business development at broad-line and startup semiconductor firms. Sankar earned a Bachelor of Applied Science and Engineering from the University of Toronto and an MBA from The Wharton School, University of Pennsylvania, and holds 5 U.S. patents in the area of VLSI architectures.