From version 3.1
edited by EANTC Bot
on 2025/03/24 17:51
Change comment: There is no comment for this version
To version 4.1
edited by Monika Leung
on 2025/03/24 18:56
Change comment: Upload new image "434507675074625537_Topology-V15.jpg", version 1.2

Summary

Details

Page properties
Author
... ... @@ -1,1 +1,1 @@
1 -xwiki:XWiki.eantc_bot
1 +xwiki:XWiki.leung
Content
... ... @@ -5,6 +5,7 @@
5 5  ----
6 6  
7 7  === Editor's Note ===
8 +
8 8  {{id name="editors_note"/}}Welcome to the latest edition of the EANTC multi-vendor interoperability test report! It has been an honor for us to gather the leading network equipment manufacturers again - Arista Networks, Arrcus, Calnex Solutions, Ciena, Ericsson, H3C Technologies, Huawei Technologies, Juniper Networks, Keysight, Microchip, Nokia, and Ribbon Communications participated. In February, a team of more than 70 engineers from vendors and EANTC conducted an intense three-week test event at our lab in Berlin.
9 9  
10 10  This year, **we split the report into a paper-based overview and a much more extensive online version. Our website has full details of all test results. Please scan the QR codes, enter the URLs displayed on this and the following pages, or browse EANTC's website** to access the individual report areas. The online report is more detailed than the printed report in previous years!
... ... @@ -12,8 +12,7 @@
12 12  This year, we have continued our coverage of multi-vendor interoperability of segment routing, EVPN, time synchronization, network management, and orchestration. Meanwhile, these technologies have matured considerably, albeit to different degrees (see more details in each report section). Once we had built a baseline of interoperable services between participating vendors, we focused on the latest innovations in each area - cloud data center and data center interconnect services, and advanced 5G/6G transport use case tests.
13 13  
14 14  Segment Routing over IPv6 (SRv6) has evolved from a challenger to a major pillar of our transport protocol tests. Its advantages - a single, scalable transport solution spanning from data centers through access and aggregation to the core network - are convincing. The number of vendors participating in this test area is stable. In parallel, Segment Routing over MPLS (SR-MPLS) remains a stronghold for service provider networks because it works on many routers without needing hardware upgrades. For the foreseeable future, SRv6 and SR-MPLS will coexist. Likewise, VXLAN maintains its role in data center (DC) networks, supported by standard switch hardware and simple to manage for DC operators. Consequently, vendors tested gateways between SR-MPLS, VXLAN, and SRv6 again.
15 -
16 -EVPN (Ethernet VPN Service) is the workhorse carrying all the end-to-end application data across all segment routing and VXLAN transport networks. We validated extension operational improvements and focused on interworking EVPN gateways between VXLAN and SR-MPLS. Additionally, we helped vendors expand the interoperability of many of the core EVPN service attributes.
16 +\\EVPN (Ethernet VPN Service) is the workhorse carrying all the end-to-end application data across all segment routing and VXLAN transport networks. We validated extension operational improvements and focused on interworking EVPN gateways between VXLAN and SR-MPLS. Additionally, we helped vendors expand the interoperability of many of the core EVPN service attributes.
17 17  
18 18  5G networks continue to evolve, with packetized transport becoming the de facto standard across all segments. Transformation particularly occurs in fronthaul networks, where Ethernet-based transport solutions substitute traditional CPRI-based connectivity, for example, in the O-RAN Alliance architecture. The migration enables greater flexibility and cost savings; at the same time, it introduces stringent transport network performance and reliability requirements. Critical parameters such as ultra-low latency, minimal packet delay variation (jitter), and precise timing synchronization must be maintained to ensure optimal radio access network operation. Our interoperability tests reflected multi-vendor, real-world deployment conditions: We evaluated x-haul SRv6 environments with parallel PTP traffic, secure PTP transport using MACsec, constraint-based routing based on policies, and more.
19 19  
... ... @@ -22,23 +22,28 @@
22 22  I sincerely thank all the participating vendor teams for their strong commitment to this year's test over many months, through an intense three-week hot staging at EANTC in Berlin and culminating in the public live display in Paris. It is an honor for EANTC to see that our interoperability efforts contribute a little to the success of many service provider, enterprise, and government networks worldwide - enabling a wide choice of innovative solutions and a lively global market. The EANTC team hopes that this test report provides new insights and proof points for network operators.
23 23  
24 24  === EANTC's Mission ===
25 +
25 25  EANTC provides vendor-neutral testing to validate the interoperability, performance, robustness, and security of network solutions, platforms, and applications. Since 1991, we have been committed to transparent and reproducible assessments, helping the industry ensure standards compliance and minimize operational risks. With over 33 years of experience, we accelerate technology development and enhance the stability of vendor solutions, enabling reliable and high-quality network deployments.
26 26  
27 27  === Working Process ===
29 +
28 28  Preparations for the MPLS & SDN interoperability event began in September 2024, with initial discussions on test areas and potential test cases held across multiple technical calls with all participating vendors. These sessions focused on reviewing test case details, exploring new testing ideas, and aligning with the latest standards to ensure the test plans are up-to-date with industry advancements.
29 29  The Hot Staging was held in Berlin in the latter half of February, where the latest hardware and software arrived from all over the world, ready for testing. Two weeks of intensive testing, detailed discussions, and quick problem-solving led to excellent outcomes for all participants vendors.
30 30  EANTC engineers carefully observed and validated each test combination, following strict procedures and predefined test steps. This report includes only those results that were consistently submitted, logged, and verified by EANTC specialists to ensure accuracy and avoid misinterpretations or false positives
31 31  
32 32  === Test Area Selection ===
35 +
33 33   EANTC defines the test areas in collaboration with participating vendors to cover all key aspects of service provider networks. Vendors contribute new test cases, often with input from IETF RFC and draft editors within their teams. Given the extensive scope, we prioritize test cases that receive broad implementation support.
34 34  Our primary focus is on multi-vendor testing, so test cases implemented by a single vendor are generally excluded. However, an exception is made if a previously validated multi-vendor test fails during hot staging, leaving only one vendor with a working, standards-compliant implementation. In such cases, we recognize their commitment and report the result.
35 35  
36 36  === Interoperability Test Results ===
40 +
37 37   As always, this test report highlights only the successful test combinations (those that passed), along with the corresponding vendor and device names. In this context, the term "tested" refers to multi-vendor interoperability tests.
38 38  Test combinations that did not pass are not shown in the diagrams but are referenced anonymously in the report to illustrate the current state of the industry. From our experience, vendors swiftly address interoperability issues following our tests, and we believe it's more productive to support their efforts to improve rather than penalize them for testing new solutions. Maintaining confidentiality is key to encouraging vendors to bring their latest—often still in beta—solutions to the table, ensuring a secure space for testing and learning.
39 39  The Test Results will be presented live at the 26th edition of the MPLS & SRv6 AI Net World Congress. For over 20 years, we have been showcasing interoperability tests at Upperside's conferences.
40 40  
41 41  === Participating Vendors and Devices ===
46 +
42 42  {{id name="vendors_devices"/}}Here is the list of the devices that the participating vendors installed and tested during the Hot Staging. In some cases, multiple fixed configurations of the same product families were tested, to explore different interface types or hardware options.
43 43  
44 44  {{container cssClass="tc-role-table"}}
... ... @@ -101,13 +101,11 @@
101 101  {{/container}}
102 102  
103 103  === Test Equipment ===
109 +
104 104  The Spirent TestCenter generated unicast, multicast and broadcast traffic in the EVPN SR-MPLS and EVPN VXLAN test beds.
105 -
106 -The Calnex Solutions Paragon-Neo PAM4 and Paragon-Neo S measured the time error between PTP-synchronized devices and analyzed synchronization signals for SyncE-synchronized devices while emulating a Telecom Grandmaster (T-GM) and Telecom Slave Clock (T-SC). The Calnex Solutions Sentinel measured the 1PPS time error for PTP-synchronized devices.
107 -
108 -The Keysight Time Sync Analyzer measured the time error between PTP-synchronized devices, analyzed synchronization signals for SyncE-synchronized devices and emulated a T-GM, T-SC, Distributed Unit (DU), and Radio Unit (RU).
109 -
110 -The Keysight IxNetwork acted as a traffic generator and network emulator, emulated the IGP underlay and BGP VPN overlay, and provided a flexible packet builder.
111 +\\The Calnex Solutions Paragon-Neo PAM4 and Paragon-Neo S measured the time error between PTP-synchronized devices and analyzed synchronization signals for SyncE-synchronized devices while emulating a Telecom Grandmaster (T-GM) and Telecom Slave Clock (T-SC). The Calnex Solutions Sentinel measured the 1PPS time error for PTP-synchronized devices.
112 +\\The Keysight Time Sync Analyzer measured the time error between PTP-synchronized devices, analyzed synchronization signals for SyncE-synchronized devices and emulated a T-GM, T-SC, Distributed Unit (DU), and Radio Unit (RU).
113 +\\The Keysight IxNetwork acted as a traffic generator and network emulator, emulated the IGP underlay and BGP VPN overlay, and provided a flexible packet builder.
111 111  
112 112  === [[EVPN>>doc:.EVPN.WebHome]] ===
113 113  
... ... @@ -120,32 +120,30 @@
120 120  === [[Time Synchronization>>doc:.Time Synchronization.WebHome]] ===
121 121  
122 122  === Overall Physical Test Topology ===
126 +
123 123  {{id name="phys_topology"/}}
124 124  
125 125  [[Figure 1: Overall Physical Test Topology>>image:434507675074625537_Topology-V15.jpg||alt="Figure 1" width="550"]]
126 126  
127 127  === Conclusion ===
132 +
128 128  {{id name="conclusion"/}}This marks the conclusion of the 23rd edition of the EANTC interoperability test — the most comprehensive and long-standing multi-vendor interoperability testing series for service provider transport networks leveraging IP, Segment Routing, and (initially) MPLS. Thanks to our trusted, longstanding collaboration with Upperside Conferences, we successfully presented an increasingly expansive and sophisticated showcase at the MPLS & SDN World Congress in Paris once again this year.
134 +\\Test outcomes were distributed as follows:
129 129  
130 -Test outcomes were distributed as follows:
131 -
132 -*. 46% in Segment Routing
133 -*. 25% in EVPN
134 -*. 19% in time synchronization
135 -*. 10% in network management and orchestration
136 136  
137 +* 46% in Segment Routing
138 +* 25% in EVPN
139 +* 19% in time synchronization
140 +* 10% in network management and orchestration
141 +
137 137  Amid the global expansion of standalone 5G deployments, many participating vendors focused on testing advanced 5G x-Haul scenarios utilizing SRv6 and SR-MPLS data planes, including traffic engineering policies and time synchronization.
138 -
139 -The results reflect a pivotal moment for the transport networking industry: Segment Routing architectures (SR-MPLS and SRv6 with EVPN) are now robust, mature, and refined. However, the industry's focus is shifting to address the needs of cloud data centers running GenAI workloads and the demands of 5G/6G-ready mobile edge services. These use cases necessitate highly scalable, reliable, and application-specific service performance.
140 -
141 -Simultaneously, service providers must rethink how they manage network operations centers (NOCs). The manual, network management software-assisted approaches of the past no longer keep pace with growing complexity. As network architectures mature and qualified, affordable personnel is challenging to hire, the urgency for automated network operations intensifies. While automated provisioning has long been a standard for consumer and small business services, fully autonomous networks (TM Forum AN Level 4) are now on the horizon. Such networks will leverage telemetry-based analytics, automated troubleshooting, and self-optimizing capabilities.
142 -
143 -In both key areas—application-specific services for AI and automated network operations—manufacturers are making rapid progress, so far sometimes with proprietary solutions. At EANTC, we advocate for the accelerated development of standards-based, interoperable solutions. Critical questions remain: How will future transport networks support supplier-independent network expansion? Will service providers retain the ability to issue RFPs for discrete components like core, aggregation, peering, and edge routers, or will they need to embrace single-vendor ecosystems to achieve automated operations? These are critical topics that we will tackle next year.
144 -
145 -We hope our tests offer valuable insights to network operators—including carriers, service providers, MNOs, large enterprises, and government agencies—demonstrating how to design efficient architectures and achieve seamless operations in multi-vendor network environments.
143 +\\The results reflect a pivotal moment for the transport networking industry: Segment Routing architectures (SR-MPLS and SRv6 with EVPN) are now robust, mature, and refined. However, the industry's focus is shifting to address the needs of cloud data centers running GenAI workloads and the demands of 5G/6G-ready mobile edge services. These use cases necessitate highly scalable, reliable, and application-specific service performance.
144 +\\Simultaneously, service providers must rethink how they manage network operations centers (NOCs). The manual, network management software-assisted approaches of the past no longer keep pace with growing complexity. As network architectures mature and qualified, affordable personnel is challenging to hire, the urgency for automated network operations intensifies. While automated provisioning has long been a standard for consumer and small business services, fully autonomous networks (TM Forum AN Level 4) are now on the horizon. Such networks will leverage telemetry-based analytics, automated troubleshooting, and self-optimizing capabilities.
145 +\\In both key areas—application-specific services for AI and automated network operations—manufacturers are making rapid progress, so far sometimes with proprietary solutions. At EANTC, we advocate for the accelerated development of standards-based, interoperable solutions. Critical questions remain: How will future transport networks support supplier-independent network expansion? Will service providers retain the ability to issue RFPs for discrete components like core, aggregation, peering, and edge routers, or will they need to embrace single-vendor ecosystems to achieve automated operations? These are critical topics that we will tackle next year.
146 +\\We hope our tests offer valuable insights to network operators—including carriers, service providers, MNOs, large enterprises, and government agencies—demonstrating how to design efficient architectures and achieve seamless operations in multi-vendor network environments.
146 146  
147 147  (% id="prev-next-links" %)
148 -||[[Next ~>>>doc:.EVPN.WebHome]]
149 +| |[[Next ~>>>doc:.EVPN.WebHome]]
149 149  )))
150 150  
151 151  (% class="col-xs-12 col-sm-5 test-report-sidebar" %)
434507675074625537_Topology-V15.jpg
Author
... ... @@ -1,1 +1,1 @@
1 -xwiki:XWiki.eantc_bot
1 +xwiki:XWiki.leung
Size
... ... @@ -1,1 +1,1 @@
1 -249.1 KB
1 +458.4 KB
Content