Results Summary and Analysis
ACX7100 and ACX7509 Functions and Performance
The CoS validation focused on demonstrating ACX7100 and ACX7509 classification, scheduling, shaping and rewriting behaviors across services leveraging a 5G xHaul infrastructure. The test scenarios included measuring latency for eCPRI Fronthaul traffic types with emulated O-RU/O-DU and O-RAN test sequences. The JVD is executed over a Seamless SR-MPLS architecture with a focus on CoS.
The topology generates reasonable multi-vector scale of Layer 2/Layer 3 connectivity services as compared with mobile network operator (MNO) and metro area network (MAN) operator expectations for real network deployments, while satisfying stringent SLA requirements.
The ACX7100 and ACX7509 meet the CoS requirements mentioned here for 5G xHaul solutions. They are capable of preserving and respecting priority forwarding throughout the network topology. These platforms are particularly well-suited for access aggregation purposes.
Across the end-to-end topology, classification and rewrite was performed on 802.1p, DSCP and EXP, as outlined in Figure 1. Table 1 summarizes these results for the included services and classification types. In dual-tag scenarios, the outer service tag is used for classification and rewrite. CoS bits can be preserved end-to-end, including for inner or outer tags.
When a port shaper is defined, applicable class of service functions adjusted to the new port speed and performed equivalently. For example, a 1G port shaper was used and transmit-rate percentages were correctly shown to be based on a 1G port speed.
Table 1 provides a summary of the executed VLAN operation scenarios and their respective outcomes. The scenarios include Untagged (UT), Single-Tagged (ST), and Dual-Tagged (DT) VLAN tag representations. Table 1 presents a condensed overview of these scenarios and their results.
| Traffic Scenario | VLAN | Ingress Classification Mapped to FC | Scheduler Honored | Rates | Codepoints Rewritten | Bits Preserved | ||||
|---|---|---|---|---|---|---|---|---|---|---|
| Fixed Classifier | TAG | 802.1p | DSCP | EXP | SH | LOW | 802.1p | DSCP | EXP | E2E |
| EVPN-VPWS | UT | √ | -- | √ | √ | -- | √ | -- | √ | √ |
| L2Circuit | UT | √ | -- | √ | √ | -- | √ | -- | √ | √ |
| pop / push | DT | √ | -- | √ | √ | -- | √ | -- | √ | √ |
| swap / swap | DT | √ | -- | √ | √ | -- | √ | -- | √ | √ |
| swap-swap / swap-swap | DT | √ | -- | √ | √ | -- | √ | -- | √ | √ |
| pop-swap / swap-push | DT | √ | -- | √ | √ | -- | √ | -- | √ | √ |
| pop-pop / push-push | DT | √ | -- | √ | √ | -- | √ | -- | √ | NA |
| push / pop | ST | √ | -- | √ | √ | -- | √ | -- | √ | √ |
| swap / swap | ST | √ | -- | √ | √ | -- | √ | -- | √ | √ |
| pop / push | ST | √ | -- | √ | √ | -- | √ | -- | √ | √ |
| swap-push / pop-swap | ST | √ | -- | √ | √ | -- | √ | -- | √ | √ |
| BA Classifier | TAG | 802.1p | DSCP | EXP | SH | LOW | 802.1p | DSCP | EXP | E2E |
| L3VPN | UT | -- | √ | √ | -- | √ | -- | √ | * | √ |
| L2VPN | UT | √ | -- | √ | -- | √ | √ | -- | * | √ |
| BGP-VPLS | UT | √ | -- | √ | -- | √ | √ | -- | * | √ |
| pop / push | DT | √ | -- | √ | -- | √ | √ | -- | √ | √ |
| swap / swap | DT | √ | -- | √ | -- | √ | √ | -- | √ | √ |
| swap-swap / swap-swap | DT | √ | -- | √ | -- | √ | √ | -- | √ | √ |
| pop-swap / swap-push | DT | √ | -- | √ | -- | √ | √ | -- | √ | √ |
| pop-pop / push-push | DT | √ | -- | √ | -- | √ | √ | -- | √ | NA |
| push / pop | ST | √ | -- | √ | -- | √ | √ | -- | √ | √ |
| swap / swap | ST | √ | -- | √ | -- | √ | √ | -- | √ | √ |
| pop / push | ST | √ | -- | √ | -- | √ | √ | -- | √ | √ |
| swap-push / pop-swap | ST | √ | -- | √ | -- | √ | √ | -- | √ | √ |
*Contact your Juniper Networks representative for details.
All listed input/output VLAN mapping operations were validated across L2Circuit, L2VPN, and EVPN services. Contact your Juniper Networks representative for an analysis explaining the results for each function.
Latency Budgets
The 5G xHaul infrastructure has specific guidelines for latency, especially in the Fronthaul segment where extremely low latency flows are crucial. The overall latency budget considers factors such as fiber length, connected devices, and transport design. O-RAN sets a maximum limit of 100µs for one-way latency from O-RU to O-DU, with each device having a latency of around ≤10µs. However, there is a growing demand from operations for device latency to be even lower, around 5-6µs. This represents a significant shift in requirements compared to earlier 4G architectures.
Table 2 provides a snapshot of the performance comparison in Fronthaul, Midhaul, and MBH for various service types on ACX7100 and ACX7509. The total latency factors in the number of hops, such as EVPN-VPWS with three hops resulting in a measurement of 15.1µs, approximately 5µs per hop in the Fronthaul segment. For a more in-depth analysis of per-device latency for ultra-low latency traffic types, refer to the Latency measurements in Table 2. If you need additional information, including specific data outputs and reports, contact your Juniper Networks representative.
| Traffic Type | Latency (µs) | Queue Type | Port Speed | Segment | DUT | Hops |
|---|---|---|---|---|---|---|
| EVPN-VPWS (eCPRI) | 15.1µs | strict-high | 10G | Fronthaul | ACX7100 | 3 |
| EVPN-VPWS (eCPRI) | 16.1µs | strict-high | 10G | Fronthaul | ACX7509 | 2 |
| EVPN-VPWS (eCPRI) | 10.6µs | strict-high | 1G Shaper | Fronthaul | ACX7100 | 3 |
| EVPN-VPWS (eCPRI) | 8.4µs | strict-high | 1G Shaper | Fronthaul | ACX7509 | 2 |
| L2Circuit | 62.3µs | Low | 100G | Midhaul | ACX7509 | 6 |
| L2Circuit | 74.6µs | Low | 10G | Midhaul | ACX7509 | 5 |
| L2Circuit | 103.5µs | Low | 1G Shaper | Midhaul | ACX7509 | 5 |
| L2Circuit | 68.2µs | Low | 100G | MBH | ACX7100 | 6 |
| L2Circuit | 81.8µs | Low | 10G | MBH | ACX7100 | 6 |
| L2Circuit | 127.4µs | Low | 1G Shaper | MBH | ACX7100 | 6 |
| L3VPN (CRIT) | 61.6µs | strict-high | 100G | Midhaul | ACX7509 | 6 |
| L3VPN | 126µs | Low | 10G | Midhaul | ACX7509 | 5 |
| L3VPN | 117.7µs | Low | 1G Shaper | Midhaul | ACX7509 | 5 |
| L3VPN | 145.9µs | Low | 10G | MBH | ACX7100 | 6 |
| L3VPN | 137.1µs | Low | 1G Shaper | MBH | ACX7100 | 6 |
In non-congestion scenarios, the priority queue provides latency performance improvement over low priority queues.