SLA typically includes three to five classes.
Real-time traffic gets fixed bandwidth allocation.
Data traffic gets variable bandwidth allocation with minimum guarantee.
Additional classes not visible to customer may exist at the edge.
A typical QoS SLA offered by most service providers typically includes three to five traffic classes; for example, a real-time traffic class, a mission-critical data traffic class, one or two other data traffic classes, and a best-effort traffic class. The SLA for the real-time traffic class should be guaranteed a fixed maximum bandwidth, while the data traffic classes should be guaranteed a minimum bandwidth. Typically, the bandwidth allocation is configured as a percentage of the interface bandwidth. Each traffic class can also have a latency, delay, jitter, and packet-loss guarantee.
Between the CE and PE, there may be additional traffic classes that are used by the service providers only. For example, there may be a management traffic class for traffic such as Telnet or Simple Network Management Protocol (SNMP) from the service provider to the service provider-managed CE routers.
If a single physical interface is only serving one customer, the SLA is typically set up per interface. To provide easy bandwidth upgrades, service providers often install a high-speed link to the customer and then offer a sub-rate access.
If a single physical interface is serving many different customers, the SLA is typically set up per-permanent virtual circuit (PVC) or per-VLAN. To provide easy bandwidth upgrades, the service provider often installs a high-speed link to the customer and then offers a sub-rate access.
Typical SLA Requirements for Voice
To meet QoS requirements for different traffic types, both the enterprise and the service provider must implement the proper QoS mechanisms to provide end-to-end QoS for the packets traversing a service provider IP network. In the figure, the enterprise headquarters and the enterprise branch office are connected to a service provider that is providing Layer 3
In this example, the service provider is providing an SLA for voice traffic with a latency of 60 ms or less, a jitter of 20 ms or less, and a packet loss of 0.5 percent or less. To meet the end-to-end QoS requirements for voice packets, the entire enterprise network must contribute less than 90 ms of delay—that is, 90 ms (enterprise network) + 60 ms (service provider network) <= 150 ms total one way delay. Similarly, jitter must be less than 10 ms—that is, 10 ms + 20 ms <= 30 ms total one way jitter. Finally, packet loss must be less than 0.5 percent—that is, 0.5 percent + 0.5 percent <= 1.0 percent total packet loss.
Was this article helpful?