QFX5240 Switches Datasheet
Download DatasheetProduct Overview
The QFX5240 Switches meet the advanced AI data center networking requirements of large-scale clusters. QFX5240 switches work with the automation—such as Juniper Apstra—to assure daily operation in AI and ML workload training and access.
QFX5240 Switches:
— deliver high-density 800GbE ports on a fixed form factor with software to provide advanced network services tuned to the specific needs of AI/ML workloads
— are a foundation of AI networks, and their low latencies ensure fast job completion time (JCT) to speed training through high GPU utilization
— help teams managing AI/ML environments realize improved economics
Product Description
Continuous evolution of AI/ML technology along with new applications are driving the next major shift in bandwidth requirement within the data center fabric. Juniper Networks® QFX5240 800GbE switch (64 port 800GbE) is a next-generation, fixed-configuration platform designed for spine, leaf, and border switch roles. The switch provides flexible, cost-effective, high-density 800GbE, 400GbE, 100GbE, and 50GbE interfaces for intra-IP fabric connectivity as well as higher density 200/400GbE NIC connectivity for AI/ML use cases. It’s 51.2Tbps unidirectional throughput meets the bandwidth requirement of AI/ML workloads and storage systems with latency in the range of 700-750ns (store and forward). Remote Direct Memory Access (RDMA) is the de-facto data transfer technology used in AI/ML workloads, and it uses Remote Direct Memory Access over Converged Ethernet v2 (ROCEv2) for transport at the network layer. QFX5240 supports ROCEv2 along with congestion management features like Priority Flow Control (PFC), explicit congestion notification (ECN), and data center quantized congestion notification (DCQCN).
The QFX5240 helps reduce the number of network nodes deployed—decreasing the total power consumption of the data center fabric and improving the carbon footprint of the data center. These improvements are possible by having different breakout options like 128x400GbE, 256x200GbE, and 320x100GbE.
AI Data Center |
|
Cloud-ready Data Center |
|
Port Options |
|
Platform Parameters |
|
Features and Benefits
AI/ML Design
Artificial intelligence puts new challenges on compute, network, and storage solutions with large models that run in parallel across many GPUs for training. These models require fast job completion time (JCT) with minimal delays for the last GPU to finish its calculations, that is, low tail latency. Architects optimize the cluster performance through rail-optimized design (Read this Juniper White Paper for more information about AI/ML cluster design). As model sizes and datasets continue to grow, designs must accommodate more GPUs in the cluster, requiring that the network seamlessly scale, without compromising performance, or introducing communication bottlenecks.
The QFX5240 meets the needs of these large-scale AI networks. The switch provides:
- 64 ports of 800GbE on a 2 U switch to reduce costs on both space and total power utilization
- Choice of connectivity with both OSFP and QSFP-DD variants of 800GbE for leaf-spine connectivity
- Advanced telemetry capabilities to support ECN/PFC counters
- Fine-grained, load-balancing capability to handle reduced flow entropy
- Automation of rail-optimized design through Apstra
Automation
Automation tools, such as Apstra, ensure the reliable set up of expansive networks with ongoing verification of the deployment along with monitoring of operations. Apstra intent-based networking delivers full Day 0 through Day 2+ capabilities for IP/EVPN fabrics with closed-loop assurance in the data center. Apstra provides a broad set of operational capabilities, with multiple built-in intent-based analytics probes, flow visibility, and analysis to ensure that the AI network is running as designed. Apstra provides a simple UI workflow to create custom intent-based analytics to capture, enrich, and visualize data from the AI network.
Monitoring
The QFX5240 supports Junos® telemetry interface, a modern telemetry streaming tool that provides performance monitoring in complex, dynamic data centers. Streaming data to a performance management system lets network administrators measure trends in link and node utilization and troubleshoot issues such as network congestion in real time.
- Application visibility and performance management by provisioning sensors to collect and stream data and analyze the application and workload flow path through the network
- Capacity planning and optimization by proactively detecting hotspots and monitoring latency and microbursts
- Troubleshooting and root cause analysis via high-frequency monitoring and correlating overlay and underlay networks
Additionally, the Junos Evolved operating system supports a robust API set to support automation through Terraform, Ansible, zero-touch provisioning (ZTP), operations and event scripts, automatic rollback, and Python scripts.
Specifications
Hardware Specifications
Parameter | QFX5240-64OD | QFX5240-64QD |
System throughput | 51.2/102.4 Tbps uni/bidirectional | 51.2/102.4 Tbps uni/bidirectional |
Max Forwarding Rate | 21.2Bpps | 21.2Bpps |
Port density | 64 ports of OSFP 800GbE | 64 ports of QSFP-DD 800GbE |
Max ports with breakout | 64 × 800GbE, 128 × 400GbE, 320 × 100GbE | 64 × 800GbE, 128 × 400GbE, 320 × 100GbE |
Dimensions (W x H x D) | 17.26 x 3.46 x 25.52 in (43.8 x 8.8 x 64.8 cm) | 17.26 x 3.46 x 25.52 in (43.8 x 8.8 x 64.8 cm) |
Rack units | 2 U | 2 U |
Weight | 22kgs (48.50lbs) fully loaded without optics | 22kgs (48.50lbs) fully loaded without optics |
Operating system | Junos OS Evolved | Junos OS Evolved |
CPU | Intel Ice Lake (4 core) | Intel Ice Lake (4 core) |
Memory | 32GB (16GBx2) of DDR4 | 32GB (16GBx2) of DDR4 |
Storage | 2x480GB | 2x480GB |
Power | Redundant (1+1) hot-pluggable 3000W AC (200 to 240V) power supplies | Redundant (1+1) hot-pluggable 3000W AC (200 to 240V) power supplies |
Cooling | Ports-to-PSU (AFO) 4 hot-pluggable fan modules | Ports-to-PSU (AFO) 4 hot-pluggable fan modules |
Total packet buffer | 165 MB | 165 MB |
Warranty | Juniper standard one-year warranty | Juniper standard one-year warranty |
Features |
Layer 2 Features |
STP—IEEE 802.1D (802.1D-2004) |
Rapid Spanning Tree Protocol (RSTP) (IEEE 802.1w); MSTP (IEEE 802.1s) |
Bridge protocol data unit (BPDU) protect |
Loop protect |
Root protect |
VLAN—IEEE 802.1Q VLAN trunking |
Routed VLAN interface (RVI) |
Static MAC address assignment for interface |
Global MAC learning disable |
Link Aggregation and Link Aggregation Control Protocol (LACP) (IEEE 802.3ad) |
IEEE 802.1AB Link Layer Discovery Protocol (LLDP) |
Layer 3 Features |
Static routing |
OSPF v2/v3 |
Filter-based forwarding |
VRRP/VRRPv3 |
IPv6 |
Virtual routers |
Loop-free alternate (LFA) |
BGP |
IS-IS |
Dynamic Host Configuration Protocol (DHCP) v4/v6 relay (stateless) |
VRF-aware DHCP |
IPv4/IPv6 over GRE tunnels |
Multicast |
Internet Group Management Protocol (IGMP) v1/v2/v3 |
Multicast Listener Discovery (MLD) v2 |
IGMP proxy, querier |
IGMP v1/v2/v3 snooping |
Intersubnet multicast using IRB interface |
MLD snooping |
Protocol Independent Multicast PIM-SM, PIM-SSM, PIM-DM, PIM-Bidir Multicast Source Discovery Protocol (MSDP) |
Quality of Service (QoS) |
L2 and L3 QoS: Classification, rewrite, queuing |
Rate limiting: - Ingress policing: 1 rate 2 color, 2 rate 3 color - Egress policing: Policer, policer mark down action - Egress shaping: Per queue, per port |
10 hardware queues per port (8 unicast and 2 multicast) |
Strict priority queuing (LLQ), shaped-deficit weighted round robin (SDWRR) |
Layer 2 classification criteria: Interface, MAC address, Ether type, 802.1p, VLAN |
Congestion avoidance capabilities: WRED, ECN |
Trust IEEE 802.1p |
Configurable shared buffer and buffer monitoring |
Congestion Notification Profile |
Priority-based flow control (PFC)—IEEE 802.1Qbb |
High Availability |
Bidirectional Forwarding Detection (BFD) |
Visibility and Analytics |
Switched Port Analyzer (SPAN) |
Remote SPAN (RSPAN) |
Encapsulated Remote SPAN (ERSPAN) |
sFlow v5 |
Junos Telemetry Interface Management and Operations |
Role-based CLI management and access |
Junos OS Evolved configuration rescue and rollback |
Image rollback |
SNMP v1/v2/v3 |
Junos OS Evolved XML management protocol |
Automation and orchestration |
Zero-touch provisioning (ZTP) |
Python |
Junos OS Evolved event, commit, and OP scripts |
Network Services |
ROCEv2 |
DCQCN, PFC, ECN |
PFC watchdog |
Dynamic load balancing (DLB) |
Configurable hash-bucket size |
Environmental Ranges
Parameter | QFX5240-64OD |
Operating temperature | 0° to 40°C @ sea level |
Storage temperature | -40° to 70°C |
Operating altitude | Up to 6000 ft/3962 m |
Relative humidity operating | 5 to 90% non-condensing |
Relative humidity nonoperating | 5 to 90% non-condensing |
Seismic | Zone 4 earthquake rating (GR-63 EQ zone 4) |
Typical power consumption | 932-Watt 100% traffic with DACs (without optics power) @ 25°C |
Ordering Information
Product SKU | Description |
QFX5240-64OD-AO | 64x800GE OSFP800, AC, Front-to-back airflow |
QFX5240-64OD-CHAS | 64x800GE OSFP800, spare chassis |
QFX5240-64QD-AO | 64x800GE QSFP-DD 800, AC, Front-to-back airflow |
QFX5240-64QD-CHAS | 64x800GE QSFP-DD 800, spare chassis |
License SKUs
S-QFX5K-C5-A1-X (X=3,5,P) | Advanced 1 Software License (X Years Subscription, X=3,5, or P for Perpetual) for QFX5240-64OD/QD |
S-QFX5K-C5-A2-X (X=3,5,P) | Advanced 2 Software License (X Years Subscription, X=1,3,5, or P for Perpetual) for QFX5240-64OD/QD |
S-QFX5K-C5-P1-X (X=3,5,P) | Premium Software License (X Years Subscription, X=1,3,5, or P for Perpetual) for QFX5240-64OD/QD |
Optics and Transceivers
Up‑to‑date information on supported optics can be found on the Hardware Compatibility Tool. The QFX5240 line of Switches supports varying port speeds at 800GbE, 400GbE, and 100GbE, with different transceiver options.
Useful Links
About Juniper Networks
Juniper Networks believes that connectivity is not the same as experiencing a great connection. Juniper's AI-Native Networking Platform is built from the ground up to leverage AI to deliver the best and most secure user experiences from the edge to the data center and cloud. Additional information can be found at Juniper Networks (www.juniper.net) or connect with Juniper on X (Twitter), LinkedIn, and Facebook.
1000792 - 04 - EN SEPTEMBER 2024