About Arista Networks
Arista Networks was founded to deliver software driven cloud networking solutions for large data center and computing environments. At the core of Arista's platform is Arista EOS, an advanced network operating system. Arista Networks products are available worldwide through distribution partners, systems integrators and resellers. The company was founded in 2004 and is based in Santa Clara, California.
Missing: Arista Networks's Product Demo & Case Studies
Promote your product offering to tech buyers.
Reach 1000s of buyers who use CB Insights to identify vendors, demo products, and make purchasing decisions.
Missing: Arista Networks's Product & Differentiators
Don’t let your products get skipped. Buyers use our vendor rankings to shortlist companies and drive requests for proposals (RFPs).
Research containing Arista Networks
Get data-driven expert analysis from the CB Insights Intelligence Unit.
CB Insights Intelligence Analysts have mentioned Arista Networks in 1 CB Insights research brief, most recently on Apr 22, 2022.
Expert Collections containing Arista Networks
Expert Collections are analyst-curated lists that highlight the companies you need to know in the most important technology spaces.
Arista Networks is included in 1 Expert Collection, including Conference Exhibitors.
Arista Networks Patents
Arista Networks has filed 563 patents.
Routers (computing), Networking hardware, Network protocols, Computer networking, Ethernet
Routers (computing), Networking hardware, Network protocols, Computer networking, Ethernet
Latest Arista Networks News
Jan 18, 2023
Reseller News Arista floats answer to the strain AI puts on networks AI Spine relies on Arista’s switches with deep packet buffers and its networking software to control heavy AI traffic, reduce latency, and eliminate packet loss. Credit: Dreamstime If networks are to deliver the full power of artificial intelligence (AI) they will need a combination of high-performance connectivity and no packet loss. The concern is that today’s traditional network interconnects cannot provide the required scale and bandwidth to keep up with AI requests, said Martin Hull, vice president of Cloud Titans and Platform Product Management with Arista Networks. Historically, the only option to connect processor cores and memory have been proprietary interconnects such as InfiniBand, PCI Express and other protocols that connect compute clusters with offloads but for the most part that won’t work with AI and its workload requirements. Arista AI Spine To address these concerns, Arista is developing a technology it calls AI Spine, which calls for switches with deep packet buffers and networking software that provides real-time monitoring to manage the buffers and efficiently control traffic. “What we are starting to see is a wave of applications based on AI, natural language, machine learning, that involve a huge ingestion of data distributed across hundreds or thousands of processors—CPUs, GPUs—all taking on that compute task, slicing it up into pieces, each processing their piece of it, and sending it back again,” Hull said. “And if your network is guilty of dropping traffic, that means that the start of the AI workload is delayed because you've got to retransmit it. And if during the processing of those AI workloads, traffic is going backwards and forwards again, that slows down the AI jobs, and they may actually fail.” AI Spine architecture Arista’s AI Spine is based on its 7800R3 Series switches, which, at the high end supports 460Tbps of switching capacity and hundreds of 40Gbps, 50Gbps, 100Gbps, or 400Gbps interfaces along with 384GB of deep buffering. “Deep buffers are the key to keeping the traffic moving and not dropping anything,” Hull said. “Some worry about latency with large buffers, but our analytics don’t show that happening here.” AI Spine systems would be controlled by Arista’s core networking software, the Extensible Operating System (EOS), which enables high-bandwidth, lossless, low-latency, Ethernet-based networks that can interconnect thousands of GPUs at speeds of 100Gbps, 400Gbps, and 800Gbps along with buffer-allocation schemes, according to a white paper on AI Spine . To help support that, the switches and EOS package creates a fabric that breaks apart packets and reformats them into uniform-sized cells, “spraying” them evenly across the fabric, according to Arista. The idea is to ensure equal access to all available paths within the fabric and zero packet loss. “A cell-based fabric is not concerned with the front-panel connection speeds, making mixing and matching 100G, 200G, and 400G of little concern,” Arista wrote. “Moreover, the cell fabric makes it immune to the 'flow collision' problems of an Ethernet fabric. A distributed scheduling mechanism is used within the switch to ensure fairness for traffic flows contending for access to a congested output port.” Because each flow uses any available path to reach its destination, the fabric is well suited to handling the “elephant flow” of heavy traffic common to AI/ML applications, and as a result, “there are no internal hot spots in the network,” Arista wrote. AI Spine models To explain how AI Spine would work, Arista’s white paper provides two examples. In the first, a dedicated leaf-and-spine design with Arista 7800s tied to perhaps hundreds of server racks, EOS’s intelligent load-balancing capabilities would control the traffic among the servers to avoid collisions. QoS classification, Explicit Congestion Notification (ECN), and Priority Flow Control (PFC) thresholds are configured on all the switches to avoid packet drops. Arista EOS’ Latency Analyzer (LANZ) determines the appropriate thresholds to avoid packet drops while keeping the throughput high and allows the network to scale while keeping latency predictive and low. The second use case, which could scale to hundreds of endpoints, connects all the GPU modes directly into the 7800R3 switches within AI Spine. The result is a fabric providing a single hop between all endpoints, driving down latency and enables a single, large, lossless network requiring no configuration or tuning,” Arista wrote. Challenges of networking AI The need for the AI Spine architecture was primarily driven by technologies and applications such as server virtualisation, application containerisation, multi-cloud computing, Web 2.0, big data, and HPC. “To optimise and increase the performance of these new technologies, a distributed scale-out, deep-buffered IP fabric has been proven to provide consistent performance that scales to support extreme ‘East-West’ traffic patterns,” Arista wrote. While it may be early for most enterprises to worry about handling large-scale AI cluster workloads, some larger environments as well as hyperscaler, financial, virtual reality, gaming, and automotive development networks are already gearing up for the traffic disruption they might cause on traditional networks. As AI workloads grow they put increasing pressure on the network for scale and bandwidth, but also for the right storage and buffer depth, having predictable latency, and handling both small packets an elephant flows, Jayshree Ullal, CEO of Arista recently told a Goldman Sachs technology gathering. “This requires a tremendous amount of engineering to make traditional Ethernet run as a back-end network to support this technology for the future and the growing use of 400G is going to add additional fuel to this development,” Ullal said.
Arista Networks Frequently Asked Questions (FAQ)
When was Arista Networks founded?
Arista Networks was founded in 2004.
Where is Arista Networks's headquarters?
Arista Networks's headquarters is located at 5453 Great America Parkway, Santa Clara.
What is Arista Networks's latest funding round?
Arista Networks's latest funding round is IPO.
How much did Arista Networks raise?
Arista Networks raised a total of $107.1M.
Who are the investors of Arista Networks?
Investors of Arista Networks include Accel, SingTel Innov8, Andy Bechtolsheim, David Cheriton, Dell Technologies Capital and 5 more.
Who are Arista Networks's competitors?
Competitors of Arista Networks include Pluribus Networks and 8 more.
Compare Arista Networks to Competitors
Arrcus provides software-powered network transformation for the interconnected world. The company's mission is to democratize networking by providing software with the most flexible consumption model at the lowest total cost of ownership (TCO). It was founded in 2016 and is based in San Jose, California.
Pica8 is focused on developing high-performance and low-cost network switches. Pica8 pioneered open networking with PicOS, the first network operating system that enables customers to easily migrate from conventional networking to software-defined networking (SDN) using commodity bare metal switches. PicOS provides extensive support for traditional switching and routing protocols and delivers SDN solutions through Pica8's adoption of Open-vSwitch (OVS). Pica8 can provide PicOS, switching hardware or both in a fully integrated package as part of an end-to-end data center SDN solution, and makes adopting SDN easy with a Starter Kit that enables deployment in hours rather than months.
Juniper Networks (NYSE: JNPR) offers products and services related to network infrastructure. The Company's Infrastructure segment provides routing and switching products that are used to control and direct network traffic from the core, through the edge, aggregation, and the customer premise equipment level. Infrastructure products include its Internet protocol (IP) routing, carrier Ethernet routing portfolio, and Ethernet switching portfolio.
Perimeter 81 provides Security Service Edge (SSE) and Zero Trust Network Access (ZTNA) solutions. Its platform offers instant deployment, unified management, and 360-degree visibility services. It serves customers operating across a wide range of industries, including healthcare and financial. It was founded in 2018 and is based in Tel Aviv, Israel.
IoTeedom is a smart city knowledge management platform.
Axis Security develops cloud-based cybersecurity technology to secure enterprise applications.
Discover the right solution for your team
The CB Insights tech market intelligence platform analyzes millions of data points on vendors, products, partnerships, and patents to help your team find their next technology solution.