Packet over SONET

Chapter Description

This sample chapter covers Packet over SONET (PoS) operation, encapsulation, protection, and convergence.

This chapter covers the following topics:

  • Evolution of voice and data networks

  • Packet over SONET applications

  • Packet over SONET operation and specification

  • Packet over SONET efficiencies

  • Packet over SONET network designs

  • Packet over SONET protection schemes

SONET is a time-division multiplexing (TDM) architecture that was designed to carry voice traffic. All traffic in SONET is broken down into slots of 64-kbps DS0 increments. A DS0 is the voice line that is typically hard-wired into homes. TDM architectures are not ideal solutions for transporting data. Cable and DSL providers have shown this with their high-data-throughput broadband offerings that do not incur the same costs as comparable TDM services would incur.

When it was discovered that computer data could be transported over telephone circuits, service providers (SPs) leveraged their existing SONET rings. SONET rings were designed and deployed to transport voice but could transport voice by breaking down the data needs into manageable pieces and transporting in 64-kbps increments. When anything less than 100 percent of a TDM circuit was used, the remainder is stuffed with arbitrary data and therefore wasted from both the customer's and SP's perspective. Frame Relay technology offered statistical multiplexing, which offered a solution to the inefficiencies of SONET-based services. Unfortunately, the designers of Frame Relay did not have quality of service (QoS) in mind with the design of the technology. Many carriers also offered zero committed information rate (CIR) services only, which guaranteed the end user absolutely no class of service (CoS). Customers found this unacceptable.

ATM offered a solution to the QoS issues of Frame Relay and offered scalability in the optical carrier (OC-n) domain. ATM relied on a fixed-size cell that is not compatible with the Ethernet technologies that most LANs employ. ATM-designed hardware includes a segmentation and reassembly (SAR) layer to translate Layer 2 frames (Ethernet, Token Ring, FDDI, and so on) into ATM cells. The SAR functionality introduced slight delays in the network, but it is prohibitively expensive and complex to design. Because of the issues associated with ATM, many vendors have not deployed OC-192 ATM interfaces at this time. There is also a concept known as cell tax with ATM deployments. ATM introduces extra overhead into each transmission because of its fixed size of 53 bytes. If a Layer 2 (Ethernet) frame does not fall on a cell boundary, the rest of the cell is padded to meet the 53-byte cell requirement. ATM cells might be efficiently multiplexed into a SONET frame, but the architecture has delays and inefficiencies that must be accounted for.

Packet over SONET (PoS) is a highly scalable protocol that overcomes many of the inefficiencies of ATM, while providing legacy support to internetworks with existing SONET architectures. PoS provides a mechanism to carry packets directly within the SONET synchronous payload envelope (SPE) using a small amount of High-Level Data Link Control (HDLC) or PPP framing.

Evolution of Voice and Data Networks

Voice and data networking is constantly evolving as the technology evolves. After the telegraph, telecommunications networks evolved to transport the spoken word. The next evolutionary step, data networks, occurred in the mid-1900s with the advent of computers. Although data networking started out small because only the largest corporations could afford computers, computers have fallen to such a low entry-level price that most people can afford to have a computer now and to be connected to the Internet. Data networks have evolved to the point that the benefits of converging voice and data networks into the same data infrastructure can no longer be ignored. SPs that have legacy SONET infrastructures can still offer customers high-speed alternatives with technologies such as PoS.

The first communications systems were mainframe computers linked to dumb terminals. The Synchronous Data Link Control (SDLC) protocol, developed by IBM, made this system possible by allowing communication between a mainframe and a remote workstation over long distances. This protocol evolved into the High-Level Data Link Control (HDLC) protocol, which provided reliable communications and the ability to manage the flow of traffic between devices. HDLC is an open industry standard protocol, whereas SDLC is an IBM proprietary protocol that must be licensed by IBM. Industry standard protocols such as TCP/IP drive the adoption and low costs of telecommunications equipment.

Cisco offered an enhanced multiprotocol version of the HDLC protocol to enable various protocols over the HDLC (High-Level Data Link Control) Layer 2 framing. This Cisco HDLC protocol is proprietary and exclusive to Cisco. The HDLC standard was loose at the time of Cisco's creation and left too much room for interpretation. When this was standardized in RFC 1619 with PPP in HDLC-like framing, Cisco's HDLC protocol was not compliant. Point-to-Point Protocol (PPP) evolved from HDLC; it offers an industry- standard way to provide multiprotocol networking abilities, as well as many enhancements such as authentication, multilink, and compression. PPP is used for many other technologies, including ISDN. HDLC and PPP are scalable to architectures with fast speeds. Figure 9-1 shows this networking evolution.

Figure 1Figure 9-1 Voice and Data Network Evolution

2. Applications for PoSPacket over SONET Operation and Specifications | Next Section