| Ensuring that the DSL customer can configure and manage the DSL equipment at their premises with a minimum of difficulty is critical to the success of any DSL service that is offered to the public. Broadband access in general and DSL in particular bring network and management complexities that formerly were the concern only of the larger enterprise to the home and small business. DSL promises a wide range of new and useful services to these customers; however, if they cannot install, configure, and manage these services easily and reliably, the services will not be utilized. Among the management issues raised by use of DSL for the customer include: -
The variety of DSL services that can be offered to the customer by a carrier or service provider. Each of these services may require different configurations and different configuration tools. For example, a home that supports both an Internet access service and a voice-over DSL service will require each of these services to be configured. -
Different providers of DSL service use different end to end architectures on their networks when they provide a DSL service. For example, some carriers use PPP over Ethernet for providing IP services to their customers, others use IP directly on AAL5. A vendor of CPE can design their equipment to work with either protocol stack. However, in order for the CPE to work with the particular carrier environment the CPE must know both the stack being used and particular parameters used by the network for that stack. It is unreasonable to expect the customer to know these technical details required to configure their devices to communicate over the DSL network. -
The DSL environment is complex and typically has divided realms of responsibility for both the network and services offered over the network. Figure 13.1 illustrates the three management realms that are typical for a DSL service. Figure 13.1. Division of realms of management control in a DSL network.  A DSL network may support multiple service providers, each of whom is responsible for the particular services that they offer. The configuration of equipment at the customer's site to support these services may be the responsibility of the service provider, the customer, or it may be shared. A user may simultaneously access services from several service providers. The carrier is responsible for configuring their network to support the end users' access to the service providers. Each service accessed may require its own path through the carrier's network and the specific parameters configured into the CPE and into equipment controlled by the service provider to use the path . However, ideally the amount of information that the three "realms" must exchange should be kept to a minimum. -
Most of the customers utilizing DSL for broadband access have little knowledge of network technology and little time for configuring or managing the network on their premises. In addition to the novelties of using a broadband access network, the customer may be installing a LAN for the first time in their home and small business. Neither the carrier nor the service provider can determine in advance what a customer will install in their home or business, yet the customer may expect assistance when they have difficulties connecting their new network to the DSL service. Both the carrier and the service provider have limited resources to provide this assistance. The following features, which address the issues raised above, are required of the configuration tools for DSL CPE: -
CPE installation by the user should not require a user manual or other specialized knowledge about the carrier's network or network configuration. -
CPE installation and setup should require a minimum of manual configuration on the part of the user or the carrier's technicians. Complete autoconfiguration without human intervention is the ultimate goal. -
CPE installation and setup should be consistent across multiple DSL access networks. Variations should be isolated from the end user. CPE from various vendors should work with many different DSL access architectures. -
CPE must work regardless of the relationships between subscriber, network, and service providers. A carrier may support multiple service providers. -
The CPE configuration methods must allow for the evolution of the services that are provided by the carriers and service providers. -
The network operator should not be required to contact the end user for information about their installed CPE. That is, the CPE should be able to automatically notify the network of its capabilities over the configuration interfaces provided by the network. This is in contrast to the situation for ISDN where the user must provide the carrier with the SPID (service profile ID) of their CPE at the time service is ordered. -
CPE and the DSL access network configurations should always be able to automatically maintain compatible settings. Reconfiguration of the ADSL access network should be allowed independently of the end user CPE. The user should never have to take action when the network is reconfigured. Conversely, rearrangements at the customer premise must never require the manual reconfiguration of the DSL access network by the carrier. -
Communications ”either manual or automated (bonding) ”between the service providers and carriers should be kept to a minimum. The DSL industry is highly "horizontally stratified" such that several separate providers may be cooperating to provide the service to the customer. In many cases, the carrier that provides the DSL service, sometimes known as the network access provider (NAP), provides only the lower portions of the protocol stack as seen by the customer. The service provider, often known as the network service provider (NSP), is responsible for the remainder of the stack and access to the remote applications. Both the NAP and NSP require the ability to interact independently with the CPE to configure their parts of the service. -
Widely deployed and standard configuration tools (e.g., PPP, SNMP, DHCP, etc.) should be used wherever possible. -
The function of existing (or future) user applications (e.g., Web browsers or e-mail programs) should not be affected by the DSL configuration methods. -
The tools to allow autoconfiguration of the CPE must be designed for various regulatory environments. |
