Ethernet Testing Overview

Ethernet, standardized by the Institute of Electrical and Electronics Engineers (IEEE), has become the world’s most widely used network technology. Evolving from being a LAN-based network technology, Ethernet is now among the most widely deployed access technology in carrier networks. Supported data rates have also increased over time, reaching speeds of up to 10 Gbit/s. Today’s standard rates range from 10 Mbit/s through to 10 Gbit/s and are supported on both electrical twisted pair or fiber-optic cables.

Figure 1. Typical 10 GigE WAN and LAN PHY applications

10 GigE introduced a departure from standard 10/100/1000 Mbit/s Ethernet, thanks to two variants that are optimized for either LAN or WAN applications. Since most 10 Gbit/s WAN links today are SONET/SDH-based, the 10 GigE specification has a second physical-layer specification allowing it to easily interact with existing SONET/SDH network elements. 10 GigE maintains the standard Ethernet frame size and format, so that Layer 3 and higher protocols are preserved. It operates over point-to-point links in full-duplex mode.

Ethernet Test Applications

The following table provides a list and explanation of the main test applications used in Ethernet testing.

Test Application	Description
BERT	The bit-error-rate test (BERT) application permits a user to send and analyze specific test patterns through Ethernet frames (standard and customized) at specific frame formats and rates. The payload (or pattern) of received frames are analyzed in detail to detect and monitor mismatch in the payload and other Ethernet error conditions. This test suite is ideal for network installation and service turn–up, as discussed below. This feature also includes ideal service-level agreement (SLA) tools such as service disruption time (SDT) and error-performance monitoring system (EPMS). SDT precisely measures service disruptions according to two different modes of operation with precision of 20 s, and EPMS collects statistics according to standardized performance parameters ITU-T G.821 and G.826.
Frame Analysis Test	The RFC 2544 test is a benchmarking methodology that specifies a series of subtests used to measure the performance of networks. EXFO’s offering concentrates on the four fundamental subtests; namely, the throughput, back-to-back (burst), frame loss, and latency measurements. These tests serve as the basis for network performance qualification and SLA validation and are discussed further in the Solutions section.
Traffic Generator	The frame analysis test application is in fact the combination of two tools—a traffic generation and a traffic reception tool.
Ethernet Analyzer	The reception tool consists of an Ethernet analyzer that collects statistics of all Ethernet data received at the analyzing port. These statistics include frame size and rate statistics, stream statistics, frame counts and rates based on user-defined filters. This tool also includes a packet jitter measurement tool that enables the measurement of jitter between consecutive test packets, a critical measurement for VoIP-enabled networks and other delay-sensitive applications.
Smart Loopback	The Smart Loopback configuration is a tool that can be used with the RFC, BERT or frame analyzer application. It enables a Packet Blazer to be set up in loopback mode with the capability to loop back received Ethernet frames, while conforming to Ethernet addressing rules. This is done by swapping addresses at all layers and re-looping them to the originating port at the same rate as received. The Smart Loopback tool also presents statistics similar to those available with the Ethernet Analyzer in order to provide statistical data relating to the traffic detected at the test port.
TCP Throughput Test	The majority of SLAs are used to characterize network performances up to the IP layer. SLA testing will mainly focus on the ability of the network to transport frames and will typically not be adequate for connection-oriented protocols such as Transmission Control Protocol (TCP). This protocol, which ensures a reliable and efficient delivery connection between two networked devices, optimizes its data flow to maximize throughput while minimizing retransmission and offers a dynamic throughput that varies with network events and errors. Often, the throughput established by applications is far from optimal for the supporting network, leading to performance inefficiencies within the application. The TCP throughput test is designed to measure performance statistics of a TCP connection between two modules configured in Local and Remote mode in an Ethernet network. This test also collects Layer 4 statistics for TCP throughput, window size, round-trip time, transmitted frames, retransmitted frames, as well as related Ethernet frame and IP packet statistics.
Higher-Layer Protocol Testing	Higher-layer data can be transmitted over Ethernet and as such can be used to gather statistics based on specific conditions. As part of the frame analysis test, a higher-layer protocol analyzer is available to collect real-time statistics on incoming streams containing IP, IP/UDP or IP/TCP payload.
Expert Mode	To help users sift through the high number of statistics that need to be analyzed for assessing the quality of the network, the EXPERT Mode tool is a monitoring tool that compares statistics against established user-defined thresholds providing a PASS or FAIL verdict. The Expert Mode tool is applicable to BERT and RFC tests.

Service-Level Agreements and Ethernet Service Acceptance Testing

A service-level agreement (SLA) is a legal contract between a service provider and a customer that specifies a required level of service. SLAs help service providers attract and retain customers, but there are also penalties associated with sub-standard service: poor customer satisfaction, increased spending on maintenance and, often, direct financial payouts. SLAs typically specify maximum downtime, mean-time-to-repair (MTTR) when outages occur, and minimum performance criteria.

Figure 2. Flow chart of Ethernet service acceptance testing

Ethernet service-acceptance testing involves specific tests to ensure that SLA validation can be performed on either part of the network (end-to-core) or on all of it (end-to-end) using either BERT over Ethernet or RFC 2544 benchmarking (see below for more details on both test).

Figure 3. End-to-end, end-to-core, and remote testing

Network Installation and Service Turn-Up

Because the transparent transport of Ethernet over physical media is becoming a common service, Ethernet is increasingly carried across a variety of Layer 1 media (e.g., 10Base-FL, 100Base-FX, 1000Base-LX) over longer distances. There is therefore a growing need to certify Ethernet transport on a bit-per-bit basis. This can be done using bit-error-rate testing (BERT).

BERT uses test patterns encapsulated into an Ethernet frame, making it possible to go from a frame-based error measurement to a BER measurement. These test patterns can be pseudo-binary random sequences (PRBS) test signals or user-defined test patterns. The robustness of optical network elements can be tested through Ethernet BERT testing using xPATS (CRPAT, CSPAT, CJTPAT, short and long CRTPAT) while turning up service over dark fibers, PON or DWDM networks.

Ethernet BERT provides the bit-per-bit error-count accuracy required for the acceptance testing of physical-medium transport systems. BERT over Ethernet should be used when Ethernet is carried transparently over the following Layer 1 media. Using BERT over Ethernet Layer 2 can be used to test the error handling capabilities of network elements like switches, as a user can inject errors within the payload or in the FCS sequence of the Ethernet frame.

Performance Validation/Qualification

TCP Performance Validation – With the wide deployment of Ethernet, service providers are now offering Ethernet connections to most of their customers. Usually, the service provider will provide an SLA based on Ethernet bandwidth availability which guarantees the throughput at Layer 2. Unfortunately, TCP throughput will not necessarily be equal to its under-layer (Ethernet) performance since it will be affected by many factors such as client traffic policing and aggregation. Client traffic policing refers to bandwidth grooming techniques used to limit the flow of traffic from clients to the service provider’s network, while aggregation refers to the practice of oversubscribing a network in order to take advantage of the bursty nature of Ethernet data transmission. Other conditions such as network congestions and frame losses will have a significant impact on the TCP throughput and must be taken into account when evaluating TCP performance.

The TCP throughput application is used to determine the performance of TCP/IP traffic transmission by measuring statistics for the TCP throughput, TCP window size, total number of TCP frames transmitted and retransmitted as well as the round-trip time. These measurements require the use of two units that emulate a TCP connection between a local and a remote (similar to a client-server connection) over an Ethernet network.

Ethernet Performance Validation – The Internet Engineering Task Force (IETF) has put together a test methodology to address the issues of performance verification at the Layer 2 and 3 levels. RFC 2544, Benchmarking Methodology for Network-Interconnect Devices, specifies the requirements and procedures for testing throughput (performance availability), latency (transmission delay), back-to-back frames (link burstability), and frame loss (service integrity).

When these measurements are performed, they provide a baseline for service providers to define SLAs with their customers. They enable service providers to validate the quality of the service delivered and can provide them with a tool to create value-added services that can be measured and demonstrated to customers. For example, these tests provide performance statistics and commissioning verification for virtual LAN (VLAN), virtual private networks (VPNs) and transparent LAN services (TLS), all of which use Ethernet as an access technology. The SLA criteria defined in RFC 2544 can be precisely measured using specialized test instruments. The performance verification is usually done once the installation is complete. The measurements are taken when the network is out of service to make sure that all parameters are controlled.