“`html

Arista 7800R3 Datasheet

In the ever-evolving landscape of data center infrastructure, the demand for unprecedented performance, scalability, and operational simplicity continues to accelerate. As businesses transition to cloud-native architectures and embrace bandwidth-intensive applications like AI, high-performance computing (HPC), and real-time analytics, the underlying network fabric must evolve to support these stringent requirements. Traditional networking approaches often struggle to keep pace, introducing bottlenecks, complexity, and increased operational costs.

The Arista 7800R3 Series of modular switches is designed to directly address these challenges, delivering a purpose-built platform capable of scaling to meet the demands of the largest hyperscale data centers and demanding service provider networks. These switches are not merely faster versions of previous generations; they represent a fundamental architectural advancement engineered for deterministic performance, high density, and exceptional power efficiency. This technical document will delve into the key features, architecture, specifications, and ordering information for the Arista 7800R3 Series, providing a comprehensive overview for network professionals seeking to understand its capabilities and potential applications.

Product Overview Arista 7800R3 Series

The Arista 7800R3 Series stands as a paramount example of purpose-built modular data center switches engineered for the most demanding network environments. Designed from the ground up, these switches offer industry-leading system throughput scaling to an impressive 460 Terabits per second (Tbps), meeting the capacity requirements of the largest data centers and service provider networks. Their core function lies in providing a highly scalable, low-latency, and deterministic network foundation capable of handling diverse traffic patterns, including those generated by modern AI and HPC workloads.

These modular switches are defined by several key technological pillars. Firstly, their system scale and performance are exceptional, offering up to 96 billion packets per second forwarding rate and a remarkable 14.4 Tbps capacity per slot. This enables configurations with up to 576 wire-speed 400G ports, alongside comprehensive support for 100G and 200G modes, with future readiness for 800G. Furthermore, the inclusion of a substantial deep packet buffer of up to 24 GB per line card is critical for absorbing traffic bursts and preventing packet loss, a common issue in high-speed networks. Another notable highlight is the support for 400G ZR and ZR+ optics, enabling long-haul data center interconnect (DCI) capabilities directly from the switch.

At the heart of the 7800R3 series is its fully scheduled lossless architecture. This isn’t a mere marketing term; it represents a fundamental shift in how traffic is handled within the switch fabric. Utilizing cell-spraying technology, the fabric achieves 100% efficiency by distributing traffic uniformly across all fabric links, effectively mitigating the impact of “elephant” and “mice” flows often seen in diverse workloads. The implementation of fully scheduled credit control and per-port Virtual Output Queuing (VOQ) meticulously eliminates head-of-line blocking, ensuring fair and deterministic packet delivery with extremely low latency – under 4 microseconds for 64-byte frames. This architectural approach is paramount for applications sensitive to jitter and packet loss.

High hardware availability is another cornerstone of the 7800R3 design. The platform incorporates dual-input grid redundant power supplies with configurable N+N redundancy options, ensuring power continuity. Redundancy extends to crucial components including 1+1 supervisor redundancy and fabric module redundancy, allowing for graceful degradation and zero performance loss in the event of a module failure. Combined with front-to-rear airflow for optimized cooling and energy-efficient power consumption figures (as low as 16W per 100G and 25W per 400G port typically), the 7800R3 is engineered for continuous, reliable operation in demanding data center environments.

Beyond core switching capabilities, the 7800R3 series excels in cloud-grade routing. It supports robust features for secure internet peering, including MACsec, IPsec, and VXLANsec encryption. The platform is also a key enabler for next-generation EVPN services, essential for modern network virtualization, 5G, and multi-access edge computing (MEC) deployments. Traditional carrier core transport protocols like LDP, RSVP-TE, and SR-TE are supported with high availability mechanisms such as Fast Reroute (FRR) and Topology Independent Loop-Free Alternate (TI-LFA). Furthermore, the 7800R3 integrates advanced timing capabilities based on IEEE 1588 (PTP) and offers open programmable APIs (JSON-RPC, NETCONF) for enhanced provisioning, telemetry, and network automation.

Advanced provisioning and monitoring are facilitated through Arista’s CloudVision platform, providing network-wide orchestration and workflow automation. Integrated tools like LANZ offer precise microburst detection, while Zero Touch Provisioning (ZTP) simplifies deployment. Accelerated sFlow provides granular visibility into network traffic without impacting performance. The resilient control plane, powered by multi-core hyper-threaded x86 CPUs with ample DRAM and flash storage, ensures stable and responsive management. Arista’s Extensible Operating System (EOS), the common software platform across all Arista products, provides stateful fault containment and repair, full access to the Linux shell, and extensibility through various scripting languages and APIs, simplifying network administration and enabling rapid innovation.

Product Model and Configuration Comparison

The Arista 7800R3 Series comprises a family of modular switches distinguished primarily by their chassis size and slot capacity, offering flexibility to match diverse deployment requirements within the data center and service provider networks. Each chassis supports a common set of supervisor, fabric, and line card modules, ensuring a consistent operational experience across the family while providing scalable density and throughput. The available chassis options cater to different scales, from smaller deployments to the largest hyperscale environments.

The series includes four distinct chassis sizes: the 7804R (4 slots), 7808R (8 slots), 7812R (12 slots), and the high-density 7816R and 7816LR (both 16 slots). The number of line card slots directly correlates with the maximum port density a system can achieve, enabling configurations with an extensive number of 100G and 400G interfaces. For instance, the 16-slot chassis can support up to 576 wire-speed 400G ports when fully populated with high-density 400G line cards. This modularity allows organizations to right-size their network infrastructure while maintaining a unified hardware and software architecture.

Fabric module configurations also vary by chassis, directly impacting the overall switching fabric capacity. The smaller 7804R and 7808R chassis typically utilize 6 fabric module slots, while the larger 7812R uses 6, and the 7816R utilizes 12 fabric module slots to support its higher throughput requirements. The 7816LR chassis uses 6 Fabric modules. These fabric modules operate in an active-active redundant configuration, providing graceful performance degradation in the event of a failure, unlike architectures where a single fabric module failure can significantly impact throughput or connectivity. This level of built-in redundancy at the fabric layer is crucial for maintaining high availability in mission-critical data center applications.

Power supply options are consistent across the chassis models, with support for both AC and DC power feeds. The systems are designed for grid redundancy, where each power supply unit (PSU) can draw power from dual input circuits. The chassis provide sufficient power supply slots to support N+N redundancy, ensuring continuous operation even with the failure of multiple PSUs or complete power feeds. This power architecture, combined with front-to-rear optimized airflow, contributes significantly to the system’s excellent power efficiency and reliability, key considerations in large data center deployments where operational costs and uptime are paramount.

All 7800R3 chassis are designed to accommodate any combination of the 7800R3 and 7800R3A series line cards. This interoperability provides unparalleled flexibility. Customers can deploy a mix of 100G and 400G interfaces within the same chassis, and even mix different line card types (e.g., OSFP and QSFP-DD) where supported. This flexibility allows for seamless migration paths and efficient utilization of chassis capacity based on ever-evolving port speed and density needs. The line cards themselves offer various port counts and interface types, including copper and optical options, along with features like FlexRoute for large-scale routing tables and integrated security capabilities like TunnelSec on select models.

In essence, the 7800R3 series’ modular design, coupled with consistent component architecture and comprehensive redundancy options across the chassis sizes, provides a highly adaptable platform. This allows organizations to build scalable and resilient networks that can evolve with their data center growth and changing application demands, all while simplifying management through a single operating system, Arista EOS.

Hardware Architecture Specifications

The Arista 7800R3 Series is architected for high performance and availability, featuring a multi-component modular design. This includes chassis, supervisor modules, fabric modules, power supplies, and line cards. The seamless integration of these elements forms a robust platform capable of meeting the demands of the most data-intensive environments.

Central to the system’s performance is its architecture that avoids a traditional midplane. Instead, line cards connect directly to fabric modules, minimizing latency and maximizing throughput. The system boasts a full End-to-End (ETE) scheduled fabric, ensuring that traffic is distributed efficiently and fairly. This lossless architecture is critical for applications like AI/ML training and high-performance computing, where even minor packet loss can significantly impact performance and job completion time.

Below is a table detailing the technical specifications for the various chassis options available within the Arista 7800R3 Series:

Chassis	DCS-7816L	DCS-7816	DCS-7812	DCS-7808	DCS-7804
Supervisor slots	2	2	2	2	2
Linecard Slots	16	16	12	8	4
Fabric Module Slots	6	12	6	6	6
Power Supply Slots (Max Power Budget)	24 (72 kW)	24 (72 kW)	18 (54 kW)	12 (36 kW)	8 (24 kW)
Physical Dimensions (HxWxD)	55.6″ x 17.4″ x 37”	55.3″ x 17.4″ x 40”	39.9″ x 17.4″ x 37″	27.7″ x17.4 x 37.0″	17.2″ x 17.4″ x 37.0”
Rack Units	32	32	23	16	10
Weight (Chassis Only)	483 lbs (219 kg)	640 lbs (291 kg)	355 lbs (161 kg)	222 lbs (101 kg)	163 lbs (74 kg)
Weight (Full System)	1523 lbs (691 kg)	1580 lbs (717 kg)	1082 lbs (490 kg)	717 lbs (325 kg)	448 lbs (203 kg)
Maximum 50G Density	4608	4608	3456	2304	1152
Maximum 100G Density	2304	2304	1728	1152	576
Maximum 400G Density	576	576	432	288	144
Maximum Throughput/PPS	460 Tbps / 96 Bpps	460 Tbps / 96 Bpps	345 Tbps / 72 Bpps	230 Tbps / 48 Bpps	115 Tbps / 24 Bpps
Max Power Consumption	37,914 W	39,264 W	27984 W	19,698 W	9,121 W
Typical Airflow	1985 cfm	1779.6 cfm	1546.6 cfm	1057.1 cfm	452.5 cfm
Max Airflow	4826.1 cfm	4251 cfm	3455.4 cfm	2358 cfm	1108.6 cfm

The electrical and environmental specifications underscore the platform’s data center readiness. With operating temperatures ranging from 0 to 40 degrees Celsius and front-to-rear airflow, the units are designed for standard data center cooling practices. Redundant and hot-swappable components, including supervisors, power supplies, fabric modules, and fans, ensure maximum uptime and ease of maintenance without requiring system shutdowns.

Power is delivered via hot-swappable 3000W AC or DC power supplies, supporting both grid and PSU redundancy. The AC units boast high efficiency with Titanium climate saver ratings, minimizing power consumption and reducing operational costs. The modular nature of the power system allows for flexible configurations to meet varying redundancy requirements.

Line cards are the interface to the network, offering flexible densities and speeds. The 7800R3 and 7800R3A line cards support a mix of 100G and 400G interfaces using industry-standard OSFP and QSFP-DD form factors. These line cards feature substantial per-port packet buffers, mitigating congestion and enabling efficient traffic flow. Select line cards also integrate capabilities like TunnelSec for hardware-accelerated wire-speed encryption, a crucial feature for secure data center interconnects and enterprise WANs.

Software Architecture Arista EOS

The operational heart of the Arista 7800R3 Series, like all Arista products, is the Arista Extensible Operating System, or EOS. This state-of-the-art network operating system is a significant differentiator, providing a consistent, programmable, and resilient software foundation across the entire Arista product portfolio, from fixed-configuration switches to high-capacity modular systems like the 7800R3. The uniformity of EOS greatly simplifies network administration and operational workflows, as network engineers can manage and automate diverse network devices using a single standard operating environment.

EOS is built upon a standard Linux kernel, which immediately provides a familiar environment for developers and network operators alike. This foundation grants full access to the Linux shell and its extensive suite of tools and utilities, enabling advanced troubleshooting, scripting, and overall system management flexibility not typically found in traditional network operating systems. The true innovation within EOS lies in its unique state sharing architecture. This architecture cleanly separates the system’s switch state (such as forwarding tables, interface status, and configuration parameters) from the protocol processing and application logic. All EOS processes, which handle specific network functions like BGP routing or OSPF, run in their own protected memory spaces. These processes communicate and share information through an in-memory transactional database.

This multi-process state sharing architecture provides several critical advantages for the Arista 7800R3 Series. Foremost among these is the foundation it provides for in-service software updates (ISSUs). Because the switch state is maintained independently of the individual processes, updates to software components can be performed without disrupting data plane forwarding, minimizing maintenance windows and maximizing network uptime. Furthermore, this architecture contributes to self-healing resiliency and stateful switchover in redundant configurations. If a protocol process encounters an issue, it can be restarted without impacting other processes or the overall forwarding path, and the state information is preserved, ensuring a seamless recovery.

The Linux foundation and open architecture of EOS also empower advanced monitoring and automation capabilities directly on the switch. Features such as Zero Touch Provisioning (ZTP) allow for automated device configuration upon initial boot-up, significantly reducing deployment time and eliminating manual errors. Tools like LANZ (Latency Analyzer) provide granular, real-time monitoring of microbursts and congestion events, giving operators early visibility into potential performance issues before they impact applications. DANZ (Data Analyzer) complements this by offering advanced traffic mirroring, filtering, and aggregation capabilities at wire speed for in-depth analysis without affecting production traffic. The ability to run Linux-based tools natively on the switch further enhances diagnostic and packet capture capabilities.

EOS’s extensibility is a key enabler for Software Defined Cloud Networking (SDCN). The platform supports open programmable APIs such as JSON-RPC, NETCONF, RESTCONF, and GNMI, along with OpenConfig YANG models. This exposes the rich state and control plane functions of the 7800R3 for programmatic access, allowing for tight integration with automation frameworks, cloud orchestration platforms (like OpenStack with OVSDB), and custom applications. Developers can leverage scripting languages like Bash and Python, or even write custom agents in C++ and GO using the EOS SDK, to tailor network behavior and integrate with their existing IT workflows and observability stacks.

In summary, Arista EOS provides a modern, open, and highly resilient software architecture that underpins the performance and capabilities of the 7800R3 Series. Its consistent image across product lines, combined with its Linux foundation and state sharing model, simplifies operations, enhances programmability, and delivers the reliability required for mission-critical data center and service provider environments.

Lossless Fully Scheduled Architecture

The Arista 7800R3 Series distinguishes itself with a fully scheduled, lossless architecture, a cornerstone technology crucial for demanding workloads like AI/ML, HPC, and ultra-low latency financial applications. Unlike traditional switch architectures that might rely on hashing or best-effort forwarding across the fabric, the 7800R3 employs a sophisticated end-to-end scheduling mechanism combined with deep buffering to ensure predictable performance and zero packet loss, even under challenging traffic conditions.

A key element of this architecture is the cell-based fabric using cell spraying. Incoming packets are segmented into fixed-size cells. These cells are then “sprayed” – distributed uniformly – across all available links within the fabric to reach the destination line card. This technique is highly effective at mitigating the impact of “elephant” flows (large, sustained transfers) that can overwhelm individual links and impact “mice” flows (small, delay-sensitive packets). By spreading the traffic across all links, the fabric achieves near 100% utilization and efficiency, preventing bottlenecks and ensuring that all traffic types are treated fairly and efficiently.

Complementing the cell-based fabric is the implementation of Virtual Output Queuing (VOQ) on every ingress port. In traditional switch architectures, if a packet destined for an egress port is blocked because that port is busy, it can cause head-of-line blocking (HOLB), preventing other packets in the same ingress queue (destined for different, available egress ports) from being forwarded. VOQ solves this by maintaining a separate queue at the ingress for every possible egress port. When an ingress port receives a packet, it is placed into the VOQ corresponding to its destination egress port. This ensures that congestion on one egress port does not impede traffic flowing to other egress ports from the same ingress. With VOQ, the ingress line card only forwards traffic when explicitly granted permission by the fabric, preventing HOLB and maximizing throughput.

The crucial coordination between ingress VOQs and the fabric is managed by a Distributed Credit Scheduling system. This system operates as a central arbiter that grants “credits” to ingress VOQs, allowing them to send cells to the fabric for specific egress ports. This scheduling process is “fully scheduled” because the fabric makes intelligent decisions about which VOQs to service based on traffic demand and available egress buffer space, rather than relying on probabilistic or less deterministic methods. This fine-grained control ensures fair allocation of fabric bandwidth and prevents any single input link from unfairly dominating fabric resources or oversubscribing the egress buffering.

Deep packet buffering, up to 24 GB per line card, works in conjunction with the scheduled architecture to absorb traffic bursts and handle transient congestion scenarios. Even with a perfectly scheduled fabric, bursts of traffic arriving simultaneously at an egress port can exceed its instantaneous forwarding capacity. The ample buffer space provides a cushion to absorb these bursts, preventing packet drops in scenarios like “in-cast” (multiple sources sending traffic to a single destination simultaneously) or speed mismatches between segments of the network. This deep buffering is particularly important for maintaining the efficiency of TCP connections, which can suffer severe performance degradation from packet loss.

The combined effect of cell spraying, VOQ, distributed credit scheduling, and deep buffering results in a truly lossless and deterministic forwarding environment. The architecture functions as if it were a single, very large, perfectly scheduled switch. This eliminates the unpredictability and packet loss associated with oversubscribed or less sophisticated architectures, providing the foundation for building reliable and high-performance networks crucial for supporting the most demanding modern applications. The measurable impact is ultra-low latency, even under heavy load, and the assurance that critical traffic will reach its destination without loss due to internal switching contention.

High Availability Features

For data centers and service provider networks, continuous availability is not a luxury but a fundamental requirement. The design of the Arista 7800R3 Series places a strong emphasis on high availability, incorporating multiple levels of redundancy and resilience to ensure uninterrupted network operation. This focus on fault tolerance spans hardware components, power systems, and the software architecture, preventing single points of failure and enabling graceful recovery from disruptive events.

At the hardware level, the 7800R3 features extensive redundancy for all critical system components. Supervisor modules, which handle the control plane and management functions, are deployed in a 1+1 redundant configuration. If the active supervisor experiences an issue, the standby supervisor can seamlessly take over control, minimizing disruption. The fabric modules, which provide the switching interconnect between line cards, are also redundant and active-active. All traffic is sprayed across the operational fabric modules. In the event of a fabric module failure, the remaining modules continue to forward traffic, albeit with a potential reduction in maximum system capacity, offering a graceful performance degradation rather than a complete outage.

Powering the modular chassis is a robust and highly available power system. The 7800R3 supports multiple power supply units (PSUs) in a distributed configuration. These PSUs can be configured for N+N redundancy, ensuring that the system can continue to operate even if multiple PSUs fail, as long as the remaining units can supply sufficient power. Furthermore, each power supply unit is designed with dual input circuits, enabling connection to separate power grids or sources. This grid redundancy protects against failures affecting a single power distribution path within facility power infrastructure, providing yet another layer of resilience.

Cooling is also a critical aspect of high availability, especially in densely populated, high-power systems. The 7800R3 utilizes a front-to-rear airflow pattern, which is generally optimized for data center hot-aisle/cold-aisle containment strategies. The fan system within the chassis and integrated into the fabric modules is also designed for N+1 redundancy. High-performance, individually serviceable fan modules provide dynamic temperature control, increasing or decreasing speed based on system load and temperature to maintain optimal operating conditions. If an individual fan module fails, the system can continue to cool effectively using the remaining fans, with alerts generated to notify operators of the need for replacement.

Beyond hardware redundancy, Arista’s EOS software plays a vital role in enhancing system availability. As discussed previously, the stateful architecture of EOS enables in-service software updates (ISSUs), allowing operators to deploy software patches or upgrades without interrupting data plane forwarding. This minimizes planned downtime for maintenance. EOS also includes self-healing capabilities, such as Stateful Fault Containment (SFC) and Stateful Fault Repair (SFR). These features allow the operating system to detect and isolate faults within individual processes or components, and in many cases, automatically restart or repair the affected element without impacting system operation or dropping traffic. This proactive approach to fault management significantly improves system stability and reduces the need for manual intervention.

Taken together, the combination of redundant, hot-swappable hardware components, a resilient power and cooling infrastructure, and the self-healing capabilities of Arista EOS ensures that the 7800R3 Series provides a highly available foundation for critical network services. This comprehensive approach to high availability is essential for maintaining the uptime and performance required by modern data centers and service provider networks, where even brief outages can have significant financial and operational consequences.

Cloud Grade Routing Capabilities

The Arista 7800R3 Series is not just a high-performance data center switch; it is also a powerful platform for implementing Cloud Grade Routing, a set of capabilities vital for modern service provider networks, data center interconnects (DCI), and network peering applications. These routing features extend beyond basic Layer 3 forwarding, incorporating advanced protocols, large-scale routing tables, security mechanisms, and automation hooks necessary for building highly scalable, resilient, and programmable routing infrastructures.

A fundamental requirement for cloud-grade routing is the ability to handle a massive number of routes. The 7800R3, particularly with the K-series line cards, leverages Arista’s innovative FlexRoute technology. FlexRoute utilizes a patented algorithmic approach to optimize the use of forwarding table resources, allowing the system to support the full internet routing table in hardware, scaling to over 5 million IPv4 or IPv6 routes depending on configuration. This capacity is crucial for Internet peering points, core routing functions, and complex service provider architectures where maintaining a complete view of global routing information is necessary. The flexibility offered by programmable forwarding profiles allows administrators to prioritize different table types (MAC, host, routes) based on the specific role of the switch in the network.

Beyond scale, the 7800R3 supports a comprehensive suite of standard routing protocols essential for complex topologies and inter-domain routing. This includes OSPF, OSPFv3, BGP, MP-BGP (for VPN services), IS-IS, and RIPv2. Advanced BGP features such as BGP FlowSpec for traffic filtering and mitigation, BGP Monitoring Protocol (BMP) for control plane visibility, and BGP-RPKI for route origin validation enhance security and operational capabilities. For efficient load balancing and redundancy in routed networks, the platform supports up to 512-way Equal Cost Multipath (ECMP), distributing traffic evenly across multiple paths, and also offers Unequal Cost Multipath (UCMP) based on BGP communities.

Network virtualization is a cornerstone of cloud infrastructure, and the 7800R3 provides robust support for advanced VPN services. This includes scalable Layer 2 and Layer 3 VPNs over MPLS (IP-VPN, 6PE, 6vPE) and comprehensive support for BGP EVPN (Ethernet VPN) over VXLAN and MPLS. EVPN is a modern standard for supporting virtual overlay networks, enabling stretching Layer 2 domains across data centers and providing efficient Layer 3 routing between virtual networks. The 7800R3 supports key EVPN features such as multi-homing for device and link redundancy, integrated routing and bridging (IRB) for efficient inter-subnet forwarding, and L2 multicast optimization.

For service provider backbones and core networks, the 7800R3 supports established MPLS protocols including LDP, RSVP-TE, and Segment Routing (SR-TE). Segment Routing is a particularly modern and efficient traffic engineering paradigm that simplifies network control. The platform supports critical SR features like TI-LFA for fast reroute, ensuring rapid recovery from link or node failures, and ISIS FlexAlgo for flexible path calculation based on custom metrics. These capabilities enable building highly engineered, resilient, and efficient core networks.

Security is integral to cloud-grade routing. The 7800R3 provides built-in hardware acceleration for various encryption technologies. MACsec (IEEE 802.1AE) offers hop-by-hop Layer 2 encryption, crucial for securing direct links. For securing traffic across untrusted networks, the innovative TunnelSec feature on select line cards provides hardware-accelerated IPsec and VXLANsec encryption, enabling wire-speed secure tunnels for DCI and WAN connectivity without relying on external appliances. Additionally, granular Access Control Lists (ACLs) with support for large entry counts and algorithmic matching provide robust policy enforcement and traffic filtering capabilities at line rate.

The combination of massive routing scale, comprehensive routing protocols, advanced VPN and traffic engineering capabilities, integrated security, and the programmability offered by Arista EOS makes the 7800R3 series a versatile and powerful platform for a wide array of cloud-grade routing applications, from internet peering to large-scale enterprise and service provider routing domains.

AI ML and HPC Network Architecture

The unique demands of Artificial Intelligence (AI), Machine Learning (ML), and High-Performance Computing (HPC) workloads necessitate a network infrastructure vastly different from traditional enterprise or cloud environments. These compute-intensive tasks involve massive datasets and require extremely high bandwidth, ultra-low latency, and absolutely lossless connectivity between interconnected accelerators (like GPUs). The Arista 7800R3 Series is particularly well-suited for building these specialized networks, thanks to its distinct architectural advantages tailored to these stringent requirements.

Maximizing the return on investment in expensive AI/ML and HPC clusters hinges on minimizing job completion times. This, in turn, is heavily dependent on the performance of the inter-accelerator communication network. Training models can involve hundreds of thousands, even millions, of small, latency-sensitive transactions distributed across many accelerators. A traditional network with even small amounts of packet loss or unpredictable latency can dramatically slow down these distributed computations, leading to underutilized resources and wasted compute cycles. The 7800R3’s fully scheduled, lossless architecture is designed precisely to prevent these performance degradations.

As previously discussed, the combination of cell-based fabric with cell spraying, Virtual Output Queuing (VOQ), and Distributed Credit Scheduling ensures 100% internal lossless forwarding and deterministic performance. In the context of AI/ML and HPC, this means that the network fabric can sustain full line-rate traffic from all ports simultaneously without dropping packets due to internal congestion. This is critical because these workloads often generate “all-to-all” communication patterns where many sources simultaneously send data to many destinations across the network, pushing the fabric to its limits. The 7800R3’s ability to handle such demanding traffic patterns losslessly is paramount for optimal cluster efficiency.

The deep packet buffers on the 7800R3 line cards also play a vital role. While the scheduled fabric prevents loss within the switch, transient congestion can still occur at ingress or egress ports due to uneven traffic arrival rates or speed mismatches. The substantial buffer capacity of up to 24GB per line card provides ample room to absorb these microbursts and in-cast scenarios, preventing head-of-line blocking at the port level and ensuring that even challenging traffic patterns are handled without discarding data. This is especially important for protocols commonly used in HPC environments that are highly sensitive to packet loss.

Furthermore, the 7800R3 series is designed to be accelerator-agnostic. It provides a high-performance Ethernet fabric that can connect any type of compute or storage endpoint. This crucial flexibility ensures that as the AI ecosystem continues its rapid pace of innovation – with new generations of GPUs, specialized AI processors, and storage technologies emerging – the network infrastructure built on the 7800R3 can adapt without requiring a complete overhaul. The support for 100G, 200G, and 400G modes with high density allows for building extremely large two-tier leaf-spine networks capable of connecting tens of thousands of accelerators in a non-blocking fashion.

Reliability and software quality are equally critical for long-running AI training jobs. A job that takes days or weeks to complete can be wasted by a network interruption or software bug. The high hardware availability features of the 7800R3, including redundant supervisors, power supplies, and fabric modules, coupled with the self-healing and in-service update capabilities of Arista EOS, contribute to minimizing downtime and ensuring job completion without restarts. The maturity and stability of EOS, proven in large-scale cloud environments, provide a robust software foundation for these demanding applications.

In summary, the Arista 7800R3’s fully scheduled lossless architecture, deep buffering, high port density, hardware reliability, and flexible Ethernet foundation make it an ideal choice for building the high-performance, deterministic networks required by modern AI/ML and HPC clusters. It provides the critical network performance necessary to maximize the utilization and efficiency of expensive compute resources and accelerate the pace of scientific discovery and technological innovation.

Advanced Monitoring and Provisioning

Effective management, monitoring, and provisioning are essential components of operating any large-scale network, and this is particularly true in the dynamic environments where the Arista 7800R3 Series is deployed. The platform integrates tightly with Arista’s comprehensive suite of network management tools and offers advanced built-in capabilities within EOS to streamline operations, enhance visibility, and facilitate rapid deployment.

At the forefront of Arista’s management offering is CloudVision. CloudVision serves as a network-wide orchestration, automation, and telemetry platform. It extends the foundational principles of Arista EOS (state sharing, programmability) across the entire network infrastructure. CloudVision provides a single point of control and visibility, enabling network operators to manage their entire Arista network as a cohesive entity. It simplifies complex workflows, such as configuration changes, software upgrades, and device provisioning, through automated tasks and compliance checks. By providing a network-wide view of state and topology, CloudVision significantly aids in troubleshooting, capacity planning, and ensuring configuration consistency.

For initial deployment and scaling, the 7800R3 supports Zero Touch Provisioning (ZTP). ZTP automates the process of configuring a new switch when it boots up for the first time. By leveraging standard protocols like DHCP and TFTP, a new switch can automatically discover its configuration server, download its appropriate configuration file and software image, and integrate into the network without requiring manual intervention. This significantly reduces the time and potential errors associated with deploying new network devices, which is especially valuable in large data center build-outs.

Real-time network visibility and anomaly detection are crucial for maintaining performance and proactively identifying issues. Arista EOS includes integrated features like the Latency Analyzer (LANZ) and Data Analyzer (DANZ). LANZ provides precise, real-time monitoring to detect microbursts and congestion events on the network ports before they cause packet drops or impact application performance. It can identify the sources of congestion and even capture affected traffic for detailed analysis. DANZ offers advanced traffic mirroring capabilities at wire speed, supporting filtering, replication, and aggregation across multiple ports. This allows operators to capture specific traffic flows for forensic analysis or feed network monitoring tools without impacting the performance of production traffic.

Accelerated sFlow is another powerful telemetry feature available on the 7800R3. sFlow is a standard technique for sampling network traffic and exporting metadata about the flows to a collector for analysis. In high-speed environments like 100G and 400G, traditional CPU-based sFlow sampling can become a bottleneck. The 7800R3 utilizes hardware-accelerated sFlow engines integrated into the line cards, capable of supporting high sampling rates even at wire speed. This provides invaluable visibility into network traffic patterns, application usage, and potential security issues without impacting the forwarding performance of the switch.

Beyond these key features, the 7800R3 supports industry-standard management protocols like SNMP (v1, v2, v3) for integration with existing Network Management Systems (NMS). It offers multiple interfaces for management access, including a dedicated management port and a serial console. Extensive system logging and environmental monitoring capabilities provide detailed information about the switch’s health and operational status. The open nature of EOS, with access to the Linux shell and robust APIs (eAPI, NETCONF, RESTCONF), further empowers network teams to build custom automation scripts, integrate with their preferred orchestration tools (like Chef or Puppet), and develop bespoke telemetry solutions.

This comprehensive suite of advanced monitoring and provisioning capabilities, tightly integrated with Arista EOS and the CloudVision platform, provides network operators with the tools necessary to effectively manage, automate, and gain deep visibility into their high-performance network infrastructure built on the Arista 7800R3 Series, leading to reduced operational complexity and improved network agility.

Wire-speed Encryption and DCI

Securing data in transit is a growing imperative for organizations, especially for data transmitted between data centers, to cloud providers, or across unsecured network segments. Traditional methods of encryption often involve external appliances, which introduce latency, complexity, and can become performance bottlenecks at high speeds. The Arista 7800R3 Series, particularly with its TunnelSec-capable line cards, addresses this challenge by providing hardware-accelerated, wire-speed encryption directly integrated into the switch, offering enhanced security without sacrificing network performance, especially for Data Center Interconnect (DCI).

Arista’s TunnelSec technology supports industry-standard encryption protocols, including IEEE 802.1AE MAC Security (MACsec), IPsec (RFC 4303), and VXLANsec. This multi-protocol support provides flexibility to secure traffic at different layers of the network stack depending on the specific use case and deployment scenario. MACsec provides hop-by-hop encryption at Layer 2, securing the direct link between two connected devices. It is ideal for physical connections within a data center or between adjacent network nodes where link-level confidentiality and integrity are required.

For securing traffic that traverses multiple network hops, potentially over third-party infrastructure or the public internet, IPsec and VXLANsec are invaluable. IPsec operates at Layer 3, creating secure tunnels between IP endpoints. This allows for encrypting traffic between routers or firewalls across the WAN. VXLANsec extends this concept to overlay networks, providing security for VXLAN encapsulated traffic. This is particularly relevant in modern virtualized data centers and cloud environments where VXLAN is frequently used to stretch Layer 2 domains or build virtual networks. Securing VXLAN tunnels with VXLANsec ensures that traffic within these overlays remains confidential and protected as it traverses the underlay network.

A key advantage of the TunnelSec implementation on the 7800R3AM and 7800R3AK line cards is that the encryption is performed in hardware at line rate across all supported interface speeds, from 10G up to 400G. This hardware offload is crucial; attempting to perform encryption in software or on less capable hardware would significantly limit the throughput and introduce unacceptable latency, negating the performance benefits of the 7800R3 platform. With TunnelSec, encryption and decryption occur seamlessly at wire speed, ensuring that security is not a bottleneck for high-bandwidth applications.

This integrated, wire-speed encryption is particularly impactful for Data Center Interconnect (DCI). Organizations often need to connect geographically dispersed data centers for disaster recovery, business continuity, or workload migration. DCI links frequently span considerable distances and may utilize service provider networks, requiring strong encryption to protect sensitive data. The 7800R3’s ability to perform high-speed MACsec over dedicated dark fiber or high-speed IPsec/VXLANsec over IP/MPLS networks directly from the switch simplifies the DCI architecture, reduces the cost and complexity of deploying separate encryption devices, and maintains the low latency critical for synchronized applications spanning multiple sites.

The support for dense 400G DWDM optics like 400ZR and ZR+ on select 7800R3A line cards further enhances DCI capabilities. These high-power coherent optical modules allow for transporting 400GbE traffic over distances up to 120km directly from the switch port, without requiring external transponders. Combining this with integrated wire-speed TunnelSec encryption means a single 7800R3 line card can provide a secure, high-bandwidth 400G DCI link, revolutionizing the deployment model for connecting data centers and significantly reducing both capital and operational expenses associated with DCI infrastructure.

In essence, the integrated wire-speed encryption offered by Arista TunnelSec on the 7800R3 series provides a powerful and efficient solution for securing high-speed network traffic. This capability, seamlessly integrated into the switch hardware and controlled via Arista EOS, is essential for organizations looking to build secure and high-performance data center interconnects and secure their network traffic without compromising on the performance and scalability offered by the 7800R3 platform.

Ordering Information

The Arista 7800R3 Series offers a variety of chassis bundles, line cards, optional components, and licenses to configure systems tailored to specific network requirements. Below is a comprehensive list of the orderable products within this series.

System Bundles

System bundles provide a convenient way to acquire the basic chassis along with essential components like power supplies, fabric modules, and a supervisor module, often including necessary rack mounting kits.

Product Number	Product Description
DCS-7804R3-BND	Arista 7804R Chassis bundle. Includes 7804 chassis, 6x3kW D1 AC PS, 6xFM-R3, 1×7800-Sup1A (Extended Rack Kit)
DCS-7804R3-BND-DC	Arista 7804R Chassis bundle. Includes 7804 chassis, 6x3kW D2 DC PS, 6xFM-R3, 1×7800-Sup1A (Extended Rack Kit)
DCS-7804R3-BND-S2	Arista 7804R Chassis bundle. Includes 7804 chassis, 6x3kW D1 AC PS, 6xFM-R3, 2×7800-Sup1A (Extended Rack Kit)
DCS-7804R3-BND-DC-S2	Arista 7804R Chassis bundle. Includes 7804 chassis, 6x3kW D2 DC PS, 6xFM-R3, 2×7800-Sup1A (Extended Rack Kit)
DCS-7804R3-BNDS	Arista 7804R Chassis bundle. Includes 7804 chassis, 6x3kW D1 AC PS, 6xFM-R3, 1×7800-Sup1S (Secure Boot Enabled, Extended Rack Kit)
DCS-7804R3-BNDS-DC	Arista 7804R Chassis bundle. Includes 7804 chassis, 6x3kW D2 DC PS, 6xFM-R3, 1×7800-Sup1S (Secure Boot Enabled, Extended Rack Kit)
DCS-7808R3-BND	Arista 7808R Chassis bundle. Includes 7808 chassis, 8x3kW D1 AC PS, 6xFM-R3, 1×7800-Sup1A (Extended Rack Kit)
DCS-7808R3-BND-DC	Arista 7808R Chassis bundle. Includes 7808 chassis, 8x3kW D2 DC PS, 6xFM-R3, 1×7800-Sup1A (Extended Rack Kit)
DCS-7808R3-BND-S2	Arista 7808R Chassis bundle. Includes 7808 chassis, 8x3kW D1 AC PS, 6xFM-R3, 2×7800-Sup1A (Extended Rack Kit)
DCS-7808R3-BND-DC-S2	Arista 7808R Chassis bundle. Includes 7808 chassis, 8x3kW D2 DC PS, 6xFM-R3, 2×7800-Sup1A (Extended Rack Kit)
DCS-7808R3A-BND	Arista 7808R Chassis bundle. Includes 7808 chassis, 6x3kW D1 AC PS, 6xFM2-R3 1×7800-Sup1A (Extended Rack Kit)
DCS-7808R3-BNDS	Arista 7808R Chassis bundle. Includes 7808 chassis, 8x3kW D1 AC PS, 6xFM-R3, 1×7800-Sup1S (Secure Boot Enabled, Extended Rack Kit)
DCS-7808R3-BNDS-DC	Arista 7808R Chassis bundle. Includes 7808 chassis, 8x3kW D2 DC PS, 6xFM-R3, 1×7800-Sup1S (Secure Boot Enabled, Extended Rack Kit)
DCS-7812R3-BND	Arista 7812R Chassis bundle. Includes 7812 chassis, 10x3kW D1 AC PS, 6xFM-R3, 1×7816-Sup (Extended Rack Kit)
DCS-7812R3-BND-18PS	Arista 7812R Chassis bundle. Includes 7812 chassis, 18x3kW D1 AC PS, 6xFM-R3, 1×7816-Sup (Extended Rack Kit)
DCS-7812R3-BND-DC	Arista 7812R Chassis bundle. Includes 7812 chassis, 10x3kW D2 DC PS, 6xFM-R3, 1×7816-Sup (Extended Rack Kit)
DCS-7812R3-BND-S2	Arista 7812R Chassis bundle. Includes 7812 chassis, 10x3kW D1 AC PS, 6xFM-R3, 2×7816-Sup (Extended Rack Kit)
DCS-7812R3-BND-DC-S2	Arista 7812R Chassis bundle. Includes 7812 chassis, 10x3kW D2 DC PS, 6xFM-R3, 2×7816-Sup (Extended Rack Kit)
DCS-7812R3-BNDS	Arista 7812R Chassis bundle. Includes 7812 chassis, 10x3kW D1 AC PS, 6xFM-R3, 1×7816-Sup1S (Secure Boot Enabled, Extended Rack Kit)
DCS-7812R3-BNDS-DC	Arista 7812R Chassis bundle. Includes 7812 chassis, 10x3kW D2 DC PS, 6xFM-R3, 1×7816-Sup1S (Secure Boot Enabled, Extended Rack Kit)
DCS-7816R3-BND	Arista 7816R Chassis bundle. Includes 7816 chassis, 12x3kW D1 AC PS, 12xFM-R3, 1×7816-Sup (Extended Rack Kit)
DCS-7816R3-BND-DC	Arista 7816R Chassis bundle. Includes 7816 chassis, 12x3kW D2 DC PS, 12xFM-R3, 1×7816-Sup (Extended Rack Kit)
DCS-7816R3-BND-S2	Arista 7816R Chassis bundle. Includes 7816 chassis, 12x3kW D1 AC PS, 12xFM-R3, 2×7816-Sup (Extended Rack Kit)
DCS-7816R3-BND-DC-S2	Arista 7816R Chassis bundle. Includes 7816 chassis, 12x3kW D2 DC PS, 12xFM-R3, 2×7816-Sup (Extended Rack Kit)
DCS-7816R3-BNDS	Arista 7816R Chassis bundle. Includes 7816 chassis, 12x3kW D1 AC PS, 12xFM-R3, 1×7816-Sup1S (Secure Boot Enabled, Extended Rack Kit)
DCS-7816R3-BNDS-DC	Arista 7816R Chassis bundle. Includes 7816 chassis, 12x3kW D2 DC PS, 12xFM-R3, 1×7816-Sup1S (Secure Boot Enabled, Extended Rack Kit)
DCS-7816LR3-BND	Arista 7816LR Chassis bundle. Includes 7816L chassis, 12x3kW D1 AC PS, 6xFM-LR3, 1×7816-Sup (Extended Rack Kit)
DCS-7816LR3-BND-DC	Arista 7816LR Chassis bundle. Includes 7816L chassis, 12x3kW D2 DC PS, 6xFM-LR3, 1×7816-Sup (Extended Rack Kit)
DCS-7816LR3-BND-S2	Arista 7816LR Chassis bundle. Includes 7816L chassis, 12x3kW D1 AC PS, 6xFM-LR3, 2×7816-Sup (Extended Rack Kit)
DCS-7816LR3-BND-DC-S2	Arista 7816LR Chassis bundle. Includes 7816L chassis, 12x3kW D2 DC PS, 6xFM-LR3, 2×7816-Sup (Extended Rack Kit)
DCS-7816LR3-BNDS	Arista 7816LR Chassis bundle. Includes 7816L chassis, 12x3kW D1 AC PS, 6xFM-LR3, 1×7816-Sup1S (Secure Boot Enabled, Extended Rack Kit)
DCS-7816LR3-BNDS-DC	Arista 7816LR Chassis bundle. Includes 7816L chassis, 12x3kW D2 DC PS, 6xFM-LR3, 1×7816-Sup1S (Secure Boot Enabled, Extended Rack Kit)

Line Cards

A wide selection of line cards is available to populate the chassis slots, offering various port densities, speeds, and specialized features.

Product Number	Product Description
DCS-7800R3A-36P-LC	7800R3A Series 36 port 400GbE OSFP line card
DCS-7800R3A-36PM-LC	7800R3A Series 36 port 400GbE OSFP with Enh MACsec line card
DCS-7800R3AK-36PM-LC	7800R3A Series 36 port 400GbE OSFP line card with large routes (Enh MACsec)
DCS-7800R3A-36D-LC	7800R3A Series 36 port 400GbE QSFP-DD line card
DCS-7800R3A-36DM-LC	7800R3A Series 36 port 400GbE QSFP-DD with Enh MACsec line card
DCS-7800R3AK-36DM-LC	7800R3A Series 36 port 400GbE QSFP-DD line card with large routes (Enh MACsec)
DCS-7800R3A-36D2-LC	7800R3A Series 36 port 400GbE QSFP-DD line card with CPU
DCS-7800R3A-36DM2-LC	7800R3A Series 36 port 400GbE QSFP-DD with Enh MACsec line card with CPU
DCS-7800R3AK-36DM2-LC	7800R3A Series 36 port 400GbE QSFP-DD line card with large routes and CPU (Enh MACsec)
DCS-7800R3-36P-LC	7800R3 Series 36 port 400GbE OSFP wirespeed line card
DCS-7800R3-36D-LC	7800R3 Series 36 port 400GbE QSFP-DD wirespeed line card
DCS-7800R3K-36DM-LC	7800R3 Series 36 port 400GbE QSFP-DD with MACsec wirespeed line card, large routes
DCS-7800R3-48CQ-LC	7800R3 Series 48 port 100GbE QSFP wirespeed line card
DCS-7800R3-48CQM-LC	7800R3 Series 48 port 100GbE QSFP with MACsec wirespeed line card
DCS-7800R3K-48CQ-LC	7800R3 Series 48 port 100GbE QSFP wirespeed line card, large routes
DCS-7800R3-48CQ2-LC	7800R3 Series 48 port 100GbE QSFP line-card with CPU
DCS-7800R3-48CQM2-LC	7800R3 Series 48 port 100GbE QSFP with Enh MACsec line-card with CPU
DCS-7800R3-48CQMS-LC	7800R3 Series 48 port 100GbE QSFP with Enh MACsec line card
DCS-7800R3K-72Y-LC	7800R3 Series 72 port 25GbE SFP wirespeed line card, large routes

Optional Components and Spares

Various optional and spare parts are available for system maintenance, redundancy, and flexible installation.

Product Number	Product Description
DCS-7804-CH	Arista 7804 chassis, 2 Supervisor slots, 4 line card slots, 6 Fabric Module slots, AC or DC option (Extended Rack Kit)
DCS-7804R3-FM	7800R3 Series Fabric Module for 7804R3 Chassis, required for fabric slots 1-6
DCS-7808-CH	Arista 7808 chassis, 2 Supervisor slots, 8 line card slots, 6 Fabric Module slots, AC or DC option (Extended Rack Kit)
DCS-7808R3-FM	7800R3 Series Fabric Module for 7808 Chassis, required for fabric slots 1-6
DCS-7808R3-FM2	7800R3 Series Fabric Module-2 for 7808 Chassis, required for fabric slots 1-6
DCS-7812-CH	Arista 7812 chassis, 2 Supervisor slots, 12 line card slots, 6 Fabric Module slots, AC or DC option (Extended Rack Kit)
DCS-7812R3-FM	7800R3 Series Fabric Module for 7812 Chassis, required for fabric slots 1-6
DCS-7816-CH	Arista 7816 chassis, 2 Supervisor slots, 16 line card slots, 12 Fabric Module slots, AC or DC option (Extended Rack Kit)
DCS-7816R3-FM	7800R3 Series Fabric Module for 7816 Chassis, required for fabric slots 1-12
DCS-7816L-CH	Arista 7816L chassis, 2 Supervisor slots, 16 line card slots, 6 Fabric Module slots, AC or DC option (Extended Rack Kit)
DCS-7816LR3-FM	7800R3 Series Fabric Module for 7816L Chassis, required for fabric slots 1-6
DCS-7800-SUP	Supervisor module for 7800 series – 7808 and 7804 chassis
DCS-7800-SUP1A	Supervisor1A module for 7800 series – 7808 and 7804 chassis
DCS-7800-SUP1S	Supervisor1S module for 7800 series – 7808 and 7804 chassis (Secure Boot Enabled)
DCS-7816-SUP	Supervisor module for 7800 series – 7816L, 7816 and 7812 chassis
DCS-7816-SUP1S	Supervisor1S module for 7800 series – 7816L, 7816 and 7812 chassis (Secure Boot Enabled)
PWR-D1-3041-AC-BLUE	Arista PSU, ATS, 1RU, AC, 3KW, BLUE
PWR-D2-3041-DC-BLUE	Arista PSU, DUAL INPUT, 1RU, DC, 3KW, BLUE
PWR-D4-3041-AC-BLUE	Arista PSU, ATS-HVAC-HVDC, 1RU, AC, 3KW, BLUE (Worldwide with 277VAC/380VDC support)
FAN-7802-H	Spare high speed fan module for Arista 7800 Series (front-to-rear airflow)
DCS-7800-SCVR	Blank cover for 7800 supervisor slot
DCS-7800-PCVR	Blank cover for 7800 power supply slot
DCS-7800-LCVR	Blank cover for 7800 line card slot
KIT-7804	Spare accessory kit for Arista 7804. 4 post mount. (12x SAF-D-C20, 4m)
KIT-7808	Spare accessory kit for Arista 7808. 4 post mount. (16x SAF-D-C20, 4m)
KIT-7812	Spare accessory kit for Arista 7812. 4 post mount. (20x SAF-D-C20, 4m)
KIT-7816	Spare accessory kit for Arista 7816 4 post mount. (24x SAF-D-C20, 4m)
KIT-7816L	Spare accessory kit for Arista 7816L 4 post mount. (24x SAF-D-C20, 4m)
KIT-7800-RK	Spare 4 post mounting kit for Arista 7800 series (Standard Depth [23.9 to 33.65 in / 60.6 to 85.5 cm])
KIT-7800-RK-L	Spare 4 post mounting kit for Arista 7800 series (Extended Depth [32.37 to 42.12 in / 82.3 to 107 cm])
CAB-AC-20A-SG-C20-1M	Power cord, SAF-D-GRID to C20 (1m)
CAB-AC-20A-SG-C20-2M	Power cord, SAF-D-GRID to C20 (2m)
CAB-AC-20A-SG-C20-3M	Power cord, SAF-D-GRID to C20 (3m)
CAB-AC-20A-SG-C20-4M	Power cord, SAF-D-GRID to C20 (4m)

Optional Licenses

Software features and capabilities are often enabled via licenses, allowing for flexible feature sets based on deployment needs.

Product Number	Product Description
LIC-MOD-1-E	Enhanced Software License for Arista Modular switches – 4 slots (OSPF, BGP, ISIS, PIM)
LIC-MOD-2-E	Enhanced Software License for Arista Modular switches – 8 slots (OSPF, BGP, ISIS, PIM)
LIC-MOD-3-E	Enhanced Software License for Arista Modular switches – 12 slots (OSPF, BGP, ISIS, PIM)
LIC-MOD-4-E	Enhanced Software License for Arista Modular switches – 16 slots (OSPF, BGP, ISIS, PIM)
LIC-MOD-1-V	Virtualization license for Arista Modular switches – 4 slots (VM Tracer and VXLAN)
LIC-MOD-2-V	Virtualization license for Arista Modular switches – 8 slots (VM Tracer and VXLAN)
LIC-MOD-3-V	Virtualization license for Arista Modular switches – 12 slots (VM Tracer and VXLAN)
LIC-MOD-4-V	Virtualization license for Arista Modular switches – 16 slots (VM Tracer and VXLAN)
LIC-MOD-1-V2	EOS Extensions, Security and Partner Integration license for Arista Modular switches – 4 slots
LIC-MOD-2-V2	EOS Extensions, Security and Partner Integration license for Arista Modular switches – 8 slots
LIC-MOD-3-V2	EOS Extensions, Security and Partner Integration license for Arista Modular switches – 12 slots
LIC-MOD-4-V2	EOS Extensions, Security and Partner Integration license for Arista Modular switches – 16 slots
LIC-MOD-1-Z	Monitoring & provisioning license for Arista Modular switches – 4 slots (ZTP, LANZ, API, TapAgg)
LIC-MOD-2-Z	Monitoring & provisioning license for Arista Modular switches – 8 slots (ZTP, LANZ, API, TapAgg)
LIC-MOD-3-Z	Monitoring & provisioning license for Arista Modular switches – 12 slots (ZTP, LANZ, API, TapAgg)
LIC-MOD-4-Z	Monitoring & provisioning license for Arista Modular switches – 16 slots (ZTP, LANZ, API, TapAgg)
LIC-MOD-1-FLX	FLX License for Arista 4-Slot Modular – Full Routing up to 2M Routes, >24K ACL, EVPN, VXLAN, SR, Adv MPLS-LER/LSR, with TE & link/node protection
LIC-MOD-2-FLX	FLX License for Arista 8-Slot Modular – Full Routing up to 2M Routes, >24K ACL, EVPN, VXLAN, SR, Adv MPLS-LER/LSR, with TE & link/node protection
LIC-MOD-3-FLX	FLX License for Arista 12-Slot Modular – Full Routing up to 2M Routes, >24K ACL, EVPN, VXLAN, SR, Adv MPLS-LER/LSR, with TE & link/node protection
LIC-MOD-4-FLX	FLX License for Arista 16-Slot Modular – Full Routing up to 2M Routes, >24K ACL, EVPN, VXLAN, SR, Adv MPLS-LER/LSR, with TE & link/node protection
LIC-MOD-1-FLX-L	FLX-Lite License for Arista 4-Slot Modular – Full Routing Up to 256K Routes, EVPN, VXLAN, SR, base MPLS LSR (no TE or link/node protection)
LIC-MOD-2-FLX-L	FLX-Lite License for Arista 8-Slot Modular – Full Routing Up to 256K Routes, EVPN, VXLAN, SR, base MPLS LSR (no TE or link/node protection)
LIC-MOD-3-FLX-L	FLX-Lite License for Arista 12-Slot Modular – Full Routing Up to 256K Routes, EVPN, VXLAN, SR, base MPLS LSR (no TE or link/node protection)
LIC-MOD-4-FLX-L	FLX-Lite License for Arista 16-Slot Modular – Full Routing Up to 256K Routes, EVPN, VXLAN, SR, base MPLS LSR (no TE or link/node protection)
LIC-MOD-1-MACSEC	MACSEC Encryption License for Arista 4-Slot Modular – Encryption with MACsec capable Linecards
LIC-MOD-2-MACSEC	MACSEC Encryption License for Arista 8-Slot Modular – Encryption with MACsec capable Linecards
LIC-MOD-3-MACSEC	MACSEC Encryption License for Arista 12-Slot Modular – Encryption with MACsec capable Linecards
LIC-MOD-4-MACSEC	MACSEC Encryption License for Arista 16-Slot Modular – Encryption with MACsec capable Linecards
LIC-MOD-1-ENCR	Enhanced Security Encryption License for Arista 4-Slot Modular – Encryption capable Linecards, TunnelSec and MACsec
LIC-MOD-2-ENCR	Enhanced Security Encryption License for Arista 8-Slot Modular – Encryption capable Linecards, TunnelSec and MACsec
LIC-MOD-3-ENCR	Enhanced Security Encryption License for Arista 12-Slot Modular – Encryption capable Linecards, TunnelSec and MACsec
LIC-MOD-4-ENCR	Enhanced Security Encryption License for Arista 16-Slot Modular – Encryption capable Linecards, TunnelSec and MACsec

To acquire Arista 7800R3 Series products with global discount offers, please visit https://itmall.sale.

“`