Hardware

New ACX7100-48L Router (ACX Series)—In Junos OS Evolved Release 21.1R1, we introduce the ACX7100-48L, a high-performance router designed to address the emerging needs of 5G, IoT, and low-latency applications. The ACX7100-48L router enables a wide variety of cloud-inspired port speeds, making this router ideal for service aggregation, large enterprises, and top-of-rack, spine-and-leaf data center applications.

The ACX7100-48L provides several capabilities that include deep buffer and advanced timing, and a comprehensive set of features that are suited for cloud metro architectures.

With a 1-U, energy-efficient, and fixed-configuration design, the ACX7100-48L has a 4.8-Tbps-forwarding-capacity ASIC and the following port configurations:

Forty-seven SFP56 ports (ports 0 through 46) that operate at 50 Gbps when you use SFP56 transceivers, at 25 Gbps when you use SFP28 transceivers, or at 10 Gbps speed when you use SFP+ transceivers.
One SFP port (port 47) that operates at 25 Gbps when you use SFP28 transceivers or at 10 Gbps when you use SFP+ transceivers.
Six double-density QSFP56 (QSFP56-DD) ports (ports 48 through 53) that operate at a default speed of 400 Gbps.

We ship the ACX7100-48L routers with redundant power supply modules (PSMs). You can order the routers with front-to-back airflow or back-to-front airflow, and with AC or DC PSMs.

To install the ACX7100-48L router hardware and perform initial software configuration, routine maintenance, and troubleshooting, see the ACX7100-48L Hardware Guide.

Table 1 summarizes the ACX7100-48L features supported in Junos OS Evolved Release 21.1R1.

Table 1: Features on the ACX7100-48L
Feature	Description
Class of service (CoS)	Support for classification and rewrite rules of all types (Inet-Prec, DSCP, DSCP-v6, IEEE-802.1p, and IEEE-802.1ad) at the logical interface level. [See Classifiers and Rewrite Rules at the Global, Physical, and Logical Interface Levels Overview.]
Class of service (CoS)	Support for port shaping and scheduling with eight virtual output queues (VoQs) per port and two scheduling priority levels (strict-high and low). This release also provides support for multiple strict-high priority queues (round-robin scheduling), multiple low-priority queues (weighted fair queuing, or WFQ, scheduling), and default deep buffers. [See Schedulers Overview for ACX Series Routers and Shared and Dedicated Buffer Memory Pools on ACX Series Routers .]
Resiliency	Support for platform resiliency to handle failures and faults related to components such as CPU, fan trays, temperature sensors, power supply units, FPGA, and optics. Fault handling includes detecting and logging the error, raising alarms, sending SNMP traps, providing indication about the error through LEDs, self-healing, and taking components out of service. [See show system errors active.]
Chassis	Support for a built-in Routing Engine. The ACX7100-48L doesn't support pluggable or redundant Routing Engines. As a result, you cannot use GRES on this router. [See show chassis hardware.]
	Support for environmental monitoring and field-replaceable unit (FRU) management. [See show chassis hardware.]
	Support for multiple port speeds. The router has 48 SFP56 and 6 QSFP56-DD ports. As of Junos OS Release 21.1R1, you can configure 47 SFP56 ports (ports 0–46) to operate at 10-Gbps and 25-Gbps speeds only. Although these 47 ports are capable of providing 50-Gbps speed, we don’t support 50-Gbps speed on these ports in this release. You cannot configure port 47 if you enable PTP on the router. However, if you disable PTP, you can configure only 25-Gbps and 10-Gbps speeds on port 47. The six QSFP56-DD ports (ports 48–53) operate at a default speed of 400 Gbps. [See Configuring Port Speed at Port Level.]
Layer 2 features	Support for the following advanced Layer 2 features: Unqualified bridging support Bridge domain without a `vlan-id` `number` statement Bridge domain with the `vlan-id` value set to none Bridge domain with a single VLAN ID Single-learning domain MAC limiting Ethernet service types: E-Line with AC interface types port, VLAN, Q-in-Q, VLAN list, and VLAN maps E-Line E-LAN E-Access E-Transit LLDP LACP IRB interface Link aggregation group (LAG) support with the following hashing algorithms: For family `multiservice`, destination and source MAC addresses For family `inet`, Layer 3 and Layer 4 For family `inet6`, Layer 3 destination and source addresses For family `inet6`, Layer 4 destination and source ports Encapsulation types: `extended-vlan-bridge` `vlan-bridge` Q-in-Q tunneling [See Understanding Layer 2 Bridge Domains and Q-in-Q Tunneling on ACX Series.]
Layer 2 features	Support for interface-mode access, interface-mode trunk, and VLAN member configuration options. [See Bridging and VLANs.]
Layer 3 features	Support for the following Layer 3 features: Longest prefix match Exception packets handling VLAN tagging modes Neighbor solicitation Unicast RPF Interface-based routing The ACX7100-48L router also supports interior gateway protocols such as OSPF, IS-IS, RIP, and ECMP. [See Configure ICMP Features, Enabling VLAN Tagging, Neighbor solicitation, Understanding Unicast RPF (Routers), OSPF Overview, IS-IS Overview, RIP User Guide, and BGP Overview.]
Layer 3 features	Support for Layer 3 features. The ACX7100-48L supports BGP for IPv4 and IPv6. This release also includes support for protocols such as ECMP, OSPF, and IS-IS. Currently, the Layer 3 features such as Layer 3 VPN, BGP PIC, BGP LU, and BGP PIC core are not supported. [See BGP Overview.]
Multicast	Support for Layer 2 multicast-related features, including support for IGMP and Multicast Listener Discovery (MLD) snooping. You can configure IGMP snooping with IGMPv1, IGMPv2, and IGMPv3, which includes support for the following: IGMP snooping in a bridge domain IGMP snooping with IRB interface configured in a bridge domain MLD snooping in a bridge domain MLD snooping with IRB interface configured in a bridge domain [See IGMP Snooping Overview.]
Routing policy	Unicast reverse path forwarding (unicast RPF) support for IPv4 and IPv6. You can reduce the impact of denial-of-service (DoS) attacks for IPv4 and IPv6 interfaces by configuring unicast RPF. You can use unicast RPF to determine the source of attacks and reject packets from unexpected source addresses on interfaces. However, we do not support unicast RPF checking for: Transit packets exiting a tunnel source interface Asymmetrical routing [See Understanding Unicast RPF (Routers).]
Firewall filters	Support for firewall filters and policers. You can configure firewall filters with packet match conditions for the `bridge domain`, `IPv4`, and `IPv6` families. This release introduces support for the following match conditions in the egress direction for IPv4 and IPV6 traffic: For IPv4: ttl tcp-flags For IPv6: destination-ip next-header traffic-class hop-limit L4 ports icmp-type/code tcp-flags The ACX71000-48L also supports firewall filter actions that include `count`, `discard`, `log`, `syslog`, and `policer`. [See Overview of Firewall Filter Match Conditions and Actions on ACX Series Routers.]
Service applications	Support for RFC 2544-based benchmarking tests. We support only the Layer 3 reflector function for these tests. The support is further restricted to: `family inet` only IPv4 source and destination addresses Software-based reflection, with a maximum frame rate for the traffic flows of 1000 kbps You use the RFC 2544 benchmarking tests to measure and demonstrate the service-level agreement (SLA) parameters before service activation. The tests measure throughput, latency, frame loss rate, and number of back-to-back frames. You can configure these tests at the `[edit services monitoring rfc2544]` hierarchy level. [See RFC 2544-Based Benchmarking Tests Overview.]
Software installation and upgrade	Support for either WAN interfaces or management interfaces to automatically download and install the appropriate software and the configuration file on your device during the ZTP bootstrap process. [See Zero Touch Provisioning.]
Software installation and upgrade	Support for secure-boot implementation based on the UEFI 2.4 standard. [See Software Installation and Upgrade Guide.]
Timing and synchronization	Support for Synchronous Ethernet with Precision Time Protocol (PTP) Transparent Clock. The transparent clock measures the residence time of PTP packets as the packets pass through the router. The network load and device architecture might result in queuing or buffering delays. These delays are the main source of packet delay variation in the router. Transparent clock adds the residence time into the correction field of the PTP packet. The client or boundary clocks can determine this resident time while they receive the PTP packet from the upstream transparent clock router. The client clock can estimate and remove these delays from the offset computation and reduce the packet’s jitter effects. In syntonized transparent clock, the transparent clock requires physical layer frequency based on the ITU-T G.8262/.1 standard. Synchronous Ethernet configuration is mandatory for enabling syntonized transparent clock. Use the `show protocols ptp` and `show ptp global-information` commands to verify the PTP transparent clock configuration status. To enable PTP transparent clock on your router, configure `syntonized-e2e-transparent` statement at the `[edit protocols ptp]` hierarchy level. See [Understanding Transparent Clocks in Precision Time Protocol and show ptp global-information.]

To view the hardware compatibility matrix for optical interfaces, transceivers, and DACs supported on ACX7100-48L, see the Hardware Compatibility Tool.