ArcOS: Powering 5G and Next-Gen Hyperscale Cloud Computing with 400G

January 23, 2019 | Devesh Garg

Over the past year, I have been fortunate to talk to a number of thought leaders and CXOs across both enterprise and service providers in an effort to understand the challenges they face within their respective businesses. One consistent issue at the top-of-mind for these executives is how to rapidly maximize the business impact of their digital transformation strategies while lowering their overall TCOs. This includes leveling the playing field for their business (providers, retail, healthcare, logistics, etc.) while competing against a new wave of technology-centric hyperscalers.

It’s well established that an insatiable demand for data exists and is being accelerated by the exponential growth in rich digital media (e.g., video streaming, gaming, AR/VR, etc.), distributed storage/content, and edge computing. New SAAS-based business models, new means of delivering superior end-customer SLA-based experiences and global competitive pressures have dramatically increased the agility and immediacy demands on the internet backbone and the cloud edge.

Application workloads are no longer monolithic entities within centralized data centers but are rapidly becoming independent, modular microservices that are increasingly distributed, and essentially, more mobile across hybrid cloud environments. The network is the key infrastructure element that connects everything, sees everything.

Usher in the Era of Faster. Smarter. Better Networks

This necessitates a building-out of higher-capacity network fabrics that interconnect multiple data center PoDs as well as multiple data centers. A high-radix, flatter network fabric design for the distributed PoDs enables increased east-west communication between application workloads.

Transforming the network in this manner – to support higher bandwidths, lower latencies, and scale-out applications – is the key opportunity for businesses to level the playing field and compete effectively. This is why the onset of higher-density 100G and 400G options with highly-optimized power utilization rating is so exciting.

With our 400G announcement, Arrcus became the industry’s first independent network operating system vendor to help hyperscalers, service providers, and enterprises running data-intensive applications stay ahead of their network-bandwidth and scale-out demands.

400G: More Bandwidth. Enhanced Scale. Advanced Traffic Management

The ArcOS®-based high-density 100G and 400G switches from leading ODMs offer far more than higher network speeds to our customers. These solutions enable customers to build highly programmable, flatter network fabrics while providing real-time visibility into network traffic (packets, flows, and events). As evidenced by the support of Broadcom’s StrataXGS Tomahawk 3 switch silicon in record time, the ArcOS architecture offers the flexibility and elasticity needed to be deployed for other use cases across the backbone and the edge.

Our new switches provide 4X the bandwidth at a fraction of the power consumption of existing switches. In addition, ArcOS® will take full advantage of on-chip improvements to support enhanced scale and traffic management capabilities including:

  • Traffic-aware scheduling for elephant flow detection and re-prioritization
  • Dynamic load balancing to optimize ECMP performance in cloud and HPC environments
  • New high-performance shared buffer architecture to reduce tail latencies
  • Dynamic group management to improve ECMP congestion handling
  • Larger Route scale

We support QSFP-DD optical module form factor, which is backward compatible with QSFP ports, with the flexibility to support other form factors (e.g., OSFP) per market demand.

ArcOS: Network Different and Take Control of your Network

Ethernet speeds have transitioned from 10G to 25G to 40G to 100G, with each transition happening quicker than the one before. Significantly, these transitions have improved both the throughput and price-performance of data center networks. Now, with the pent-up demand among hyperscale cloud providers for higher capacities, transition to 400G is imminent but I see the technology becoming more mainstream only towards late 2019 (primarily due to the availability and cost-effectiveness of the optics). In the interim, customers can build-out with our power- and cost-efficient high-density 100G platforms. ArcOS deployed in this manner is future-proofed for 400G and beyond owing to its modern-day architecture.

Arrcus prides itself in meeting and exceeding the growing demands of customers for superior performance, higher scale, programmability, and security. Evolution to high-density 100G and 400G is the next big step in that journey. With our new portfolio of switches, we offer our customers unprecedented capacity while also delivering leading price-performance, choice, flexibility, and programmability needed for 5G and hyperscale cloud deployments.