Resilient Packet Ring Feature Guide
IEEE 802.17 Resilient Packet Ring Feature Guide
This feature guide describes how to configure the Cisco implementation of the IEEE 802.17 Resilient Packet Ring (RPR) protocol on supported Cisco routers and includes information about the benefits of the feature, supported platforms, related publications, and so on. RPR is similar but not identical to the Spatial Reuse Protocol (SRP), the underlying technology used in the Cisco Dynamic Packet Transfer (DPT) family of products. Throughout this document, this feature is referred to as RPR.
This document covers the use of the RPR feature. It does not include hardware installation and initial configuration information. Refer to the appropriate router installation and configuration note for information on how to configure the hardware and prepare it for use with RPR.
Information About RPR
Resilient Packet Ring (RPR), as described in IEEE 802.17, is a metropolitan area network (MAN) technology supporting data transfer among stations interconnected in a dual-ring configuration. This protocol is very similar to Spatial Reuse Protocol (SRP), which was designed by Cisco and implemented in Dynamic Packet Transport (DPT) products. New DPT interfaces have been designed to include the 802.17 RPR protocol and are available for multiple Cisco router platforms. This guide describes the RPR interface and how to use RPR on compliant Cisco equipment.
RPR is a high-speed MAC-layer protocol that is optimized for packet transmission in resilient ring topologies. RPR employs a ring structure using unidirectional, counter-rotating ringlets. Each ringlet is made up of links with data flow in the same direction. The ringlets are identified as ringlet 0 and ringlet 1, as shown in Figure 1. The use of dual fiber-optic rings provides a high level of packet survivability. If a station fails or fiber is cut, data is transmitted over the alternate ring.
Figure 1 Dual-Ring Structure
As shown in Figure 1, the east interface of Station 1 (S1) transmits to and receives from the west interface of Station 2 (S2). Ringlet 0 always transverses from east to west and ringlet 1 from west to east. The west span is the span on which RPR transmits on ringlet 1 and the east span is the span on which RPR transmits on ringlet 0.
RPR stations dynamically share the ring bandwidth and permit many simultaneous conversations. Spatial bandwidth reuse is possible due to the packet destination-stripping property of RPR. RPR provides efficient use of available bandwidth by allowing the destination station to remove unicast packets after they are read, thereby providing bandwidth reuse for the other stations on the RPR ring.
Figure 2 illustrates the end-to-end MAC architecture of RPR.
Figure 2 End-to-End View of MAC Architecture
While DPT and SRP uses SONET/SDH as the physical medium, IEEE 802.17 RPR has been defined to use both SONET/SDH and the layer used for Gigabit and 10 Gigabit Ethernet.
Comparison of RPR with SRP and DPT Technologies
IEEE 802.17 RPR is very similar to the Cisco-developed SRP technology, which is used in the Cisco DPT product line. Besides their different frame formats, other differences and similarities between IEEE 802.17 RPR and SRP can be summarized as follows:
•Fairness
–IEEE 802.17 RPR has a fairness algorithm that is used in the dynamic SRP-like mode suitable for routing and switching applications.
–A third priority has been added for traffic that requires guaranteed bandwidth, but that is not sensitive to latency and jitter.
•Protection
–SRP supports wrapping.
–IEEE 802.17 RPR supports systems that are capable of steering only protection.
–Cisco-implemented RPR supports both wrapping and steering for protection.
–Wrapping requires two stations to perform protection and suffers the least packet loss.
–Steering requires that every station determines the location of the failure and avoids that particular span. Steering is slower to converge in large topologies versus wrapping.
RPR Features
RPR offers the following main features:
•Addressing. Unicast, multicast, and simple broadcast data transfers are supported.
•Services. Multiple service qualities are supported. Per-service-quality flow-control protocols regulate traffic introduced by clients.
–Class A. The allocated or guaranteed bandwidth has low circumference-independent jitter.
–Class B. The allocated or guaranteed bandwidth has bounded circumference-dependent jitter. This class allows for transmissions of excess information rate (EIR) bandwidths (with class C properties).
–Class C. This class provides best-effort services.
•Efficiency. Design strategies increase effective bandwidths beyond those of a broadcast ring.
–Concurrent transmission. Clockwise and counterclockwise transmissions can be concurrent.
–Bandwidth reallocation. Bandwidths can be reallocated on nonoverlapping segments.
–Bandwidth reclamation. Unused bandwidths can be reclaimed by opportunistic services.
–Spatial bandwidth reuse. Opportunistic bandwidths are reused on nonoverlapping segments.
–Temporal bandwidth reuse. Unused opportunistic bandwidth can be consumed by others.
•Fairness. Fairness ensures proper partitioning of opportunistic traffic.
–Weighted. Weighted fairness allows a weighted fair access to available ring capacity.
–Simple. Simple fairness provides point-of-congestion flow control.
–Detailed. The (optional) multichoke fairness allows the client to selectively throttle its transmissions based on multiple congestion point indications.
•Plug-and-play. Automatic topology discovery and advertisement of station capabilities allow systems to become operational without manual intervention.
•Robustness. Multiple features support robust frame transmissions.
–Responsive. Service restoration time is less than 50 milliseconds after a station or link failure.
–Lossless. Queue and shaper specifications avoid frame loss in normal operation.
–Tolerant. Fully distributed control architecture eliminates single points of failure.
–OAM. Operations, administration, and maintenance support service provider environments.
This feature guide describes how to configure the Cisco implementation of the IEEE 802.17 Resilient Packet Ring (RPR) protocol on supported Cisco routers and includes information about the benefits of the feature, supported platforms, related publications, and so on. RPR is similar but not identical to the Spatial Reuse Protocol (SRP), the underlying technology used in the Cisco Dynamic Packet Transfer (DPT) family of products. Throughout this document, this feature is referred to as RPR.
This document covers the use of the RPR feature. It does not include hardware installation and initial configuration information. Refer to the appropriate router installation and configuration note for information on how to configure the hardware and prepare it for use with RPR.
Information About RPR
Resilient Packet Ring (RPR), as described in IEEE 802.17, is a metropolitan area network (MAN) technology supporting data transfer among stations interconnected in a dual-ring configuration. This protocol is very similar to Spatial Reuse Protocol (SRP), which was designed by Cisco and implemented in Dynamic Packet Transport (DPT) products. New DPT interfaces have been designed to include the 802.17 RPR protocol and are available for multiple Cisco router platforms. This guide describes the RPR interface and how to use RPR on compliant Cisco equipment.
RPR is a high-speed MAC-layer protocol that is optimized for packet transmission in resilient ring topologies. RPR employs a ring structure using unidirectional, counter-rotating ringlets. Each ringlet is made up of links with data flow in the same direction. The ringlets are identified as ringlet 0 and ringlet 1, as shown in Figure 1. The use of dual fiber-optic rings provides a high level of packet survivability. If a station fails or fiber is cut, data is transmitted over the alternate ring.
Figure 1 Dual-Ring Structure
As shown in Figure 1, the east interface of Station 1 (S1) transmits to and receives from the west interface of Station 2 (S2). Ringlet 0 always transverses from east to west and ringlet 1 from west to east. The west span is the span on which RPR transmits on ringlet 1 and the east span is the span on which RPR transmits on ringlet 0.
RPR stations dynamically share the ring bandwidth and permit many simultaneous conversations. Spatial bandwidth reuse is possible due to the packet destination-stripping property of RPR. RPR provides efficient use of available bandwidth by allowing the destination station to remove unicast packets after they are read, thereby providing bandwidth reuse for the other stations on the RPR ring.
Figure 2 illustrates the end-to-end MAC architecture of RPR.
Figure 2 End-to-End View of MAC Architecture
While DPT and SRP uses SONET/SDH as the physical medium, IEEE 802.17 RPR has been defined to use both SONET/SDH and the layer used for Gigabit and 10 Gigabit Ethernet.
Comparison of RPR with SRP and DPT Technologies
IEEE 802.17 RPR is very similar to the Cisco-developed SRP technology, which is used in the Cisco DPT product line. Besides their different frame formats, other differences and similarities between IEEE 802.17 RPR and SRP can be summarized as follows:
•Fairness
–IEEE 802.17 RPR has a fairness algorithm that is used in the dynamic SRP-like mode suitable for routing and switching applications.
–A third priority has been added for traffic that requires guaranteed bandwidth, but that is not sensitive to latency and jitter.
•Protection
–SRP supports wrapping.
–IEEE 802.17 RPR supports systems that are capable of steering only protection.
–Cisco-implemented RPR supports both wrapping and steering for protection.
–Wrapping requires two stations to perform protection and suffers the least packet loss.
–Steering requires that every station determines the location of the failure and avoids that particular span. Steering is slower to converge in large topologies versus wrapping.
RPR Features
RPR offers the following main features:
•Addressing. Unicast, multicast, and simple broadcast data transfers are supported.
•Services. Multiple service qualities are supported. Per-service-quality flow-control protocols regulate traffic introduced by clients.
–Class A. The allocated or guaranteed bandwidth has low circumference-independent jitter.
–Class B. The allocated or guaranteed bandwidth has bounded circumference-dependent jitter. This class allows for transmissions of excess information rate (EIR) bandwidths (with class C properties).
–Class C. This class provides best-effort services.
•Efficiency. Design strategies increase effective bandwidths beyond those of a broadcast ring.
–Concurrent transmission. Clockwise and counterclockwise transmissions can be concurrent.
–Bandwidth reallocation. Bandwidths can be reallocated on nonoverlapping segments.
–Bandwidth reclamation. Unused bandwidths can be reclaimed by opportunistic services.
–Spatial bandwidth reuse. Opportunistic bandwidths are reused on nonoverlapping segments.
–Temporal bandwidth reuse. Unused opportunistic bandwidth can be consumed by others.
•Fairness. Fairness ensures proper partitioning of opportunistic traffic.
–Weighted. Weighted fairness allows a weighted fair access to available ring capacity.
–Simple. Simple fairness provides point-of-congestion flow control.
–Detailed. The (optional) multichoke fairness allows the client to selectively throttle its transmissions based on multiple congestion point indications.
•Plug-and-play. Automatic topology discovery and advertisement of station capabilities allow systems to become operational without manual intervention.
•Robustness. Multiple features support robust frame transmissions.
–Responsive. Service restoration time is less than 50 milliseconds after a station or link failure.
–Lossless. Queue and shaper specifications avoid frame loss in normal operation.
–Tolerant. Fully distributed control architecture eliminates single points of failure.
–OAM. Operations, administration, and maintenance support service provider environments.
Comments