Next Generation SONET/SDH

Chapter 3.10 - Resilient Packet Ring

3.10   RESILIENT PACKET RING

Two known issues with SONET/SDH are resource utilization and cost. Two known
issues with Ethernet are latency and jitter. An attempt to minimize these issues in
the metro space and combine the virtues of both SONET and Ethernet has resulted
in another new protocol known as Resilient Packet Ring (RPR), which is consider
by some to be the rival to Next Generation SONET/SDH.

SONET and Ethernet are Layer 1 protocols. RPR is a Layer-2 MAC protocol
(being worked by the 802.17 RPR Working Group) that supports traditional carrierclass
features on dual (bidirectional) rings such as, QoS, restoration, protection and
resiliency, access and bandwidth fairness to customers, simplified provisioning, and
new services (broadcast and multicast). Based on this, RPR is supported by both
SONET and GbE on the physical layer.

A RPR packet consists of the header, the information payload and the trailingend
fields.

  • The header contains a destination address (DA), a source address (SA), a payload
    type (type), and a cyclic redundancy code (CRC) HEC code to detect and
    correct errors in the DA, SA, and payload type fields.
  • A payload field contains the information payload, which can be up to 1,500
    bytes, and the payload FCS to detect and correct errors in the payload.
  • The trailing-end field contains a time-to-live field (TTL), a class of service
    field (CoS), and an in/out profile indicator (IOP). The IOP advertises the
    drop eligibility of the packet. Based on the destination address and the IOP
    value, the MAC determines whether the packet should be dropped or passed
    in transit.

Thus, RPR packets are marked so that the MAC determines if a packet should be:

  • Dropped (or expressly switched) from the ring and sent to its destination
  • Passed in transit, avoiding the switching function
  • Stripped if the packet is corrupted
  • Copied and pass the copies in transit for multicasting

Express switching and in-transit passing eliminates or minimizes buffering, jitter,
and latency.

In particular, the RPR MAC offers four services: reserved, high priority, medium
priority, and low priority.

  • “Reserved” is a service similar to TDM, and idle bandwidth is not available to
    other services.
  • “High priority” is for jitter and latency-sensitive traffic.
  • “Medium priority” allows for services that require provisioned bandwidth.
  • “Low priority” allows for bandwidth negotiations (via bandwidth-notification
    messages) between the stations on the ring.

To reduce latency, RPR uses two counter-rotating rings to propagate packets. Each
station on the ring is designed such that packets not addressed to it pass in transit
without being switched. RPR can be synchronized with a stratum-1 8 kHz clock,
thus supporting SONET ring applications. RPR can be mapped in one or more OC-
3s of an OC-12 or OC-192, allowing for TDM to be mapped in the balance payload
envelope capacity. Thus, RPR also supports real-time traffic.

The RPR ring serves as a shared medium. Both rings transport traffic with different
priority; the lowest priority traffic is dropped first during failures. Each node on
the ring has visibility on ring capacity and it shares bandwidth according to a fairness
algorithm built into the RPR protocol.

RPR requires 50-ms protection switching (which is compatible with
SONET/SDH). Protection is accomplished with steering all traffic to the other ring,
or wrapping traffic (also known as looping back) at the nodes that are adjacent to
the failed link.

The network architecture of a RPR ring is illustrated in Figure 3.17.

UNLIMITED FREE
ACCESS
TO THE WORLD'S BEST IDEAS

SUBMIT
Already a GlobalSpec user? Log in.

This is embarrasing...

An error occurred while processing the form. Please try again in a few minutes.

Customize Your GlobalSpec Experience

Category: Fiber Optic Couplers
Finish!
Privacy Policy

This is embarrasing...

An error occurred while processing the form. Please try again in a few minutes.