Turnkey end-to-end broadband networks exploit ATM and Sonet technologies
Turnkey end-to-end broadband networks exploit ATM and Sonet technologies
Customer demand for high-speed digital service applications is forcing equipment and product suppliers into partnerships to structure complete fiber-optic broadband networks using ATM switching and Sonet transmission technologies
william l. (lonnie) martin
adc telecommunications inc.
Advanced fiber-optic synchronous optical network (Sonet) transmission systems and asynchronous transfer mode (ATM) switches can deliver voice, data and video transmissions worldwide at higher rates than can legacy DS-3 and DS-1 networks at 44.736 and 1.544 megabits per second, respectively. As a result, they are considered by most telecommunications equipment and product suppliers to be the key technologies of the future because they can provide the network backbone for bandwidth-intensive applications.
Technological advances in Sonet and ATM, the shift from voice to data and video network traffic, and the trend toward deregulation, as demonstrated by the government`s approval of the Telecommunications Act of 1996, have led industry analysts to predict a $2.5-billion increase in the telecommunications market during 1997 and 1998. Consequently, increased business opportunities to find niche markets and secure substantial growth have emerged for companies that supply high-speed network solutions and services.
To take advantage of this open-competition business climate, equipment manufacturers are striving to offer service providers complete turnkey network solutions. In turn, service providers want to offer reliable, flexible information transfers without having to rely on multiple vendors` equipment and the hassle of inherent compatibility and maintenance issues.
For example, two telecommunications manufacturers--ADC Telecommunications Inc. in Minnetonka, MN, and Hitachi Telecom (USA) Inc. in Norcross, GA--have formed an alliance to provide a turnkey broadband network that can deliver multimedia services. This alliance will offer complete end-to-end connectivity to integrate local area networks (LANs) into wide area networks (WANs) and provide public network access and transmission, network management and broadband connectivity products (see figure).
As enterprise networks become increasingly complex, managing numerous information-handling nodes with dedicated routing becomes difficult and expensive, especially when linking to small, remote offices across the public network. In contrast, modern switch-based networks offered by service providers can economically manage thousands of nodes for delivering voice, video and data communications.
Gaining access to these switch-based networks had been difficult because traditional multiplexers dedicated portions of the 1.554-Mbit/sec T1 access line to a specific premises-based device. In this approach, the number of devices per line is limited, and access bandwidth is typically wasted. Today`s access servers can connect and integrate all enterprise network communications equipment, such as telephones, computers, LANs and video equipment, to the services offered by network providers.
In addition, these servers are equipped to handle the daily variations in enterprise network traffic, as well as disaster recovery backup. Instead of maintaining and paying for dedicated leased lines for overflow and backup purposes, these servers can also dial up a circuit when needed and allow user-initiated bandwidth changes, customer premises equipment-initiated bandwidth changes, and recovery from line failure.
What has been needed to support this class of equipment and application requirements for services is a flexible network that provides LAN-to-WAN integration. The enterprise side has been moving ahead rapidly. What takes place on the public network side to support this LAN internetworking is equally important in establishing a successful application.
At the core of the new public network access and transmission networks are Sonet and ATM technologies. Since the 1980s, Sonet has been touted as the high-capacity, high-speed transmission mode of the future. While the initial technology was acceptable to meet the needs of basic or first-generation network applications, second-generation applications require advanced technology that offers more bandwidth across the backbone, such as handling 10-gigabit-per-second OC-192 terminals. To serve the "not-so-future" applications, network providers need to look for specific product attributes.
The most important requirement to be met in today`s evolving broadband network is, perhaps, advanced bandwidth management. Most Sonet network elements provide some basic time-slot assignment (TSA) functionality, which allows 52-Mbit/sec STS-1 traffic to be added, dropped or passed through a selected Sonet network element. In addition, VT1.5 add/drop traffic, along with a simple form of drop and continue or broadcast functionality, is supported by some manufacturers.
The TSA functionality falls short of expectations when emerging applications are considered. These applications can be addressed only with advanced Sonet architectures. The latest network elements, therefore, have incorporated a flexible time-slot interchange (TSI) operation.
This type of operation provides all the line-to-tributary assignment capabilities of TSA plus a full range of tributary-to-tributary crossconnect functions. In essence, TSI provides an "anything-to-anywhere" crossconnection capability and in many applications removes the requirement for expensive external digital crossconnect switches.
Moreover, TSI provides broadcast functionality across multiple network elements, including multiple drops within the same network element to accommodate advanced applications such as distance learning and ring interconnection. It also conserves headend ports and, ultimately, bandwidth for more-efficient and cost-effective network management.
To service next-generation applications, network providers can provide software upgrades via remote download. This technology protects against product obsolescence by facilitating upgrades as new Sonet services evolve via Flash electrically programmable read-only memory. It also helps avoid time-consuming and expensive site visits. In addition, network recovery time after power failures is less than that of volatile RAM-based products, which must wait for lengthy tape downloads from a software storage device before full operations can resume.
Other product specifications needed for servicing the spectrum of high-speed Sonet applications include:
Low speed interfaces--the network element should incorporate an advanced bus structure that allows a flexible combination of DS-1, DS-3 and OC-1 or STS-1 low- speed tributary support.
High-speed interface--the high-speed OC-3 interface should provide standards-compliant 156-Mbit/sec optical devices that support the current Sonet Phase EE APS protocol.
OAM&P interfaces--a variety of operations, administration, maintenance and provision interfaces should be available, including TBOS, TL-1/x.25, parallel telemetry and CMISE/SAN/1. A craft interface should also be provided for VT-100 terminal support along with a modem interface allowing remote dial-up access.
Synchronization interface--this interface should either receive or source fully redundant external 1.544-megahert¥building integrated timing supply input/output.
Maintenance and diagnostic fea tures--an extensive set of maintenance and diagnostic features should be included for maximum service availability and ease of repair. Detailed performance monitoring should be offered for all high- and low-speed interfaces. In addition, alarms should be available.
For large Sonet and ATM networks, network management is the primary consideration. For maximum interoperability, the network management system should operate in a standards-based program, such as Hewlett-Packard Co.`s Openview. Standards conformance is especially important when using equipment from multiple vendors.
Manufacturers are working to develop a single, comprehensive system for managing all network elements. In the meantime, Sonet and ATM management systems currently deployed should share a common platform for implementing the overall operation, administration and maintenance of the network`s elements. To simplify the management function, the system should include a graphical user interface, and communications between the management system and the network elements should employ a standards-based connection protocol such as simple network management protocol (SNMP).
ATM at the edge
Because of the increased user demand for ATM, network providers must install ATM switches in the early stages of deployment. These switches should be reliable, cost-effective and flexible to serve a diverse customer base. For these reasons, many providers are approaching ATM implementation in their networks from the outside in, or from "the edge of the network."
To maximize the benefits of ATM access at the network`s edge, providers must closely examine the characteristics of the equipment they select. Because not all ATM switches are designed for the same type of use, telephone companies need to consider several issues to ensure their network`s reliability, efficiency and quality of service for end users.
First-generation ATM switches did not encompass low-speed ATM interfaces, therefore a main concern for end users in the implementation of ATM backbone networks is the lack of DS-1 access flexibility. When implementing an ATM edge switch or concentrator, providers need to select equipment that can maintain multiple low-speed interfaces within the same unit. Interfaces that support existing and anticipated applications include ATM, switched multimegabit data service, frame relay, circuit emulation and frame-user network interface.
In addition to the challenges of supporting multiple interfaces, public network providers also must make sure their network is reliable. This can be accomplished by using equipment that offers protection and redundancy for all parts of the architecture, including the common control processor, the switching fabric, or bus, and the interfaces.
Carriers ideally want 1:n protection for low-speed interfaces, such as DS-1, and 1:1 protection for high-speed interfaces. The 1:n protection switching ensures that equipment chassis slots and the switch core or bus bandwidth are efficiently used. The 1:1 protection of high-speed interfaces becomes a critical equipment consideration that can affect the overall reliability and efficiency of the network.
From a public network provider`s viewpoint, the key implication of offering multiple DS-1 interfaces to customers is the ability to efficiently manage and aggregate the resulting traffic. Primary traffic challenges include buffering, policing and reporting. Performance tests on first-generation ATM switches with high-speed interfaces and small buffer sizes (hundreds to thousands of cells) have shown that even when standard Ethernet traffic was transmitted, cells were indiscriminately lost. By offering low-speed interfaces and optimized buffers, traffic can be effectively protected and segmented, giving network users and providers confidence that cells are not being discarded.
Once traffic is segmented, the network access device must give network managers the flexibility to take those segments, which may include constant bit rate, variable bit rate, internal network management and user data, and assign them to a queue based upon their appropriate class of service. This flexibility ensures that high-priority traffic service is maintained and not lost due to low-priority cells being transmitted instead.
When deploying ATM edge switches, service providers must also consider the meaningful differences between enterprise and public network equipment. For example, equipment racks in the telephone company world are built to accommodate units that are 23 inches wide. Enterprise standards call for equipment racks that are 19 inches wide. More important, the powering schemes between the two environments are different. In the enterprise world, powering is done through alternating current, whereas public networks use direct current. As a result, deploying an enterprise-intended edge switch in a public network may require power converters, and in some instances, might be incompatible.
Most public network facilities have visual and audible alarms to monitor the range of site equipment. These alarms enable fast, efficient equipment-failure detection. Because this detection is not a consideration in the enterprise world, customer premises switches designed for that environment typically are not equipped with such fault indicators. u
William L. (Lonnie) Martin is president of marketing for the network services division at ADC Telecommunications Inc., Minnetonka, MN.