(Accedian Networks') EtherNEWS Blog

Accedian Networks’ EtherNID®, MetroNID® and MetroNODE™ units are engineered to provide total quality of service (QoS) performance monitoring at both the Ethernet and IP layers. A combination of passive and active test technology is used to provide a comprehensive, precise, and granular end-to-end view of any chosen flow, service, network or link.

Passive monitoring provides per-flow traffic statistics for up to 60 unique services with per-second sampling, providing an unrivalled real-time view into service usage for troubleshooting, engineering and billing purposes.

Accurately measuring one-way / round trip latency and jitter, packet loss, continuity and availability is best accomplished using active testing techniques. Accedian Networks’ units employ the Ethernet OAM, ITU-T Y.1731 performance monitoring (PM) standard, as well as the more advanced Performance Assurance Agent (PAA)™ – capable of measuring performance at both layer 2 & 3 (Ethernet & IP).

These techniques conduct measurements by transmitting a sparse but regular stream of precisely time-stamped “tracer” packets within the service under test (in-band). The test packets’ headers mimic those of the application or SLA of interest (e.g. by specifying VLAN, CoS / DSCP, protocol, drop eligibility, etc.), to ensure they follow the same path, experience the same delay, and are given the same priority as the monitored service. The test packets’ role is to provide a recurring, known reference from which SLA / QoS metrics can be measured, without having a noticeable affect on the service itself (i.e. non-service affecting, non-intrusive).

While the test packets themselves are sent in-band, measurement results, test setup and unit communication are normally performed out-of-band using a separate management network or management VLAN apart from the service under test. This minimizes the amount of bandwidth associated with in-band active QoS testing.

Can You Quantify “Non-Intrusive” ?

The amount of capacity consumed by active testing techniques is normally insignificant – fractions of a 10th of a percent of the monitored circuit or service. The amount of bandwidth required depends on which methods are in use (PAA, OAM PM or both), the testing interval (frequency), the size of the packets specified, and the number of concurrent test sessions originating or terminating at a given test end-point. Up to 100 simultaneous PAA and / or OAM PM sessions are supported by 10 / 100 Mbps / GbE units, up to 1,000 sessions at 10 GbE.

Topology Considerations

The most demanding test conditions normally reflect hub-and spoke or full-mesh topologies. For example, in a typical wireless backhaul deployment, up to 100 towers may home into a single mobile switching center (MSC), with each link provisioned to support multiple priorities (e.g. real-time, best effort, management).

A typical monitoring deployment would involve a MetroNODE10GbE™ packet performance node at the MSC communicating with EtherNID® units at each tower. If PAA is used to monitor QoS at layer 3, and Y.1731 is used to monitor layer 2, each having sessions for all three service classes, the MSC would have to support 300 PAA and 300 OAM PM sessions concurrently. These tests would consume less than one percent of the 10 GbE link.

The amount consumed at each tower would be similarly insignificant, as only three PAA and three Y.1731 sessions would be required to fully monitor QoS. Assuming each tower is provisioned with 100 Mbps, test bandwidth would only amount to less than three percent of link capacity.