The IP Detail Record Initiative

There is no question the Internet has revolutionized communications with applications such as Web browsing and e-mail messaging. However, if the Internet is going to continue its growth and impact it must move beyond a Web and e-mail and network to one that supports real-time services, multicast, production applications, electronic commerce, multimedia messaging, and other advanced services. To make this “next-generation,” multi-service IP network a reality requires advances in IP security, quality of service and access technologies as well as an increase in core bandwidth. Such technology is here today, albeit in its infancy. Technology alone, however, is not enough. To make the next-generation IP services a commercial success requires robust and cost-effective business, operations and network support systems. Without them, how can a provider—

· Bill for application usage, content provided or network resources consumed?
· Quickly detect fraud?
· Authenticate and authorize service?
·Provision a voice, video or SLA-based data service?
· Support sophisticated credit structures such as prepaid and direct-debit?
· Make money?
The answer, of course, is: they can’t.

Next-generation IP support systems

The next-generation IP services discussed above have one thing in common: they must be priced by usage in order to be managed profitably. As such, bringing these services to market will require usage-sensitive systems for provisioning, data collection, billing, reporting and network management. Developing support systems optimized for advanced IP services is a considerable challenge. Fortunately, equipment, mediation, billing and network management vendors have developed excellent components that address these issues. In particular, today’s network elements and mediation systems can generate high-resolution usage information on a per-transaction basis. Likewise, network management and provisioning systems can configure the network in real time. Finally, billing, reporting and other business support systems optimized for IP services are beginning to appear in the market.

The remaining challenge to the IP industry is to develop interfaces between these various support systems such that they can easily exchange usage, provisioning and control information. In most networks, the fundamental unit of data exchange is the usage record, which circuit-switched vendors commonly refer to as a call detail record (CDR). Unfortunately, the CDR standards for telephony are inadequate for generic IP services, nor can an appropriate standard be found in the Internet community or IP standards bodies. Because an open, standard IP usage detail record does not exist, IP network elements and support systems cannot communicate cost-effectively, thus creating a barrier to bringing next-generation services to market.

The missing element: An open detail record

Without a standard IP usage detail record, network elements and support systems are forced to develop proprietary interfaces and mechanisms to capture usage. If there were only a few network vendors or a few IP services, then this might not be a problem. However, IP networks have a particularly heterogeneous infrastructure, comprising a wide-range of technologies and vendors. Likewise, it is envisioned that IP networks will provide a wide range of services—everything from e-mail to high-end video to “micro-transactions”—each having distinct usage parameters. With each network element and service provider capturing usage independently, the following issues arise:
· Inconsistent metrics: Network elements and mediation systems capture inconsistent usage characteristics for a given service. For example, consider IP telephony. One gateway might record only basic information such as duration, called number and calling number, whereas another gateway might include time of day, CODEC used and quality of service provisioned, as well as characterize the network service delivered (e.g., in terms of drop rates, delay and delay variation). Further, even if the same two network elements capture the same data, the meaning might not be the same. For example, does amount of data exchanged include retransmission, protocol or other overhead?

·Proprietary access: The data structures that encapsulate the usage metrics are distinct for each network element and service type. Thus, a special filter needs to be written to interpret each usage type and network element type. Likewise, special interfaces will be needed, since each network element might have a different access protocol and handle security differently. Finally, each network element might capture usage information in real time or batch the data. This affects the degree of synchronization between the support systems and actual usage, as well as how long a given system might retain usage data.

·Inconsistent provisioning: Each provider defines services differently, so each has different provisioning requirements. For example, with service level agreements (SLAs) or QoS-based services, a carrier might or might not guarantee availability, mean time to repair, drop rates, throughput or delay. Regardless of how a given SLA is defined, the network elements and management systems must be able to provision, collect and represent the data consistently. Otherwise, a carrier will be bound to a given vendor and unable to mix and match network elements based on price and functionality.

The case for a standard IP usage detail record

The current proprietary environment requires that each support system develop, at considerable expense, interfaces to each network element as well as interpret the meaning of the metrics provided. Likewise, vendors and mediation systems must use their best judgment as to what usage data to collect, based on their vision of what providers might wish to bill or report on. Given the complexity of the Internet services and the diversity of BSS/OSS requirements, such guesswork is futile. In fact, many vendors distinguish themselves on these proprietary interfaces alone for competitive advantage in the market.

A closed, proprietary interface is not the answer. The IP industry needs a standard definition of IP-based services and related usage metrics that can be a common reference model for billing and other BSS systems, service provisioning, reporting and usage collection. This common standard would:
· Eliminate the need to build costly interfaces to each network element or mediation system or between support systems.
· Provide standard usage metrics for vendors to collect and export for a given service.
· Specify the precise meaning of each usage metric.
· Specify how long network elements need to retain usage information.
· Facilitate provisioning.
· Allow billing systems to expose usage metrics in their rating interface consistently.

The IP detail record initiative

Characterizing the usage metrics for IP service is a daunting effort. Even the most basic IP service, e-mail, has a number of meaningful cost components for each transaction (size, distance, time of day, delivery options, disk storage for queued mail and POP3 queries). Real-time services such as voice and video substantially increase the difficulty, as they require a far more complex set of metrics to characterize the service and the quality delivered. The purpose of the IP detail record (IPDR) initiative is to define the essential elements of data exchange between network elements, operation support systems and business support systems. It will provide the foundation for open, carrier-grade IP support systems that enable next-generation IP networks to operate efficiently and cost-effectively. Specific goals include:
· Define an open, flexible record format (the IPDR record) for exchanging usage information between network elements or network management systems and mediation, operation support, billing or other business support systems required by an IP-based service provider.
· Define essential parameters for any IP transaction.
· Provide an extension mechanism so network elements and support systems can exchange optional usage metrics for a particular service.
· Provide a repository for defined IPDR formats

Work is underway to define prototype IPDR formats for common services, and represent the usage in state-of-the-art encapsulation techniques such as XML. After the preliminary work is complete, the IPDR formats will be taken public in an industry-wide effort to reach consensus on IP service definitions, core usage metrics and optional usage metrics. The IPDR formats will be published in a repository and evolved over time as needed.

IPDR working group structure

The IPDR effort involves a collaborative team of leading Internet system integrators, network providers, and billing, mediation and equipment vendors. The IPDR steering committee establishes the working group’s direction, agenda and review process. The representation on the steering committee is balanced across the various Internet industry segments identified above. General (non-voting) IPDR membership is open to the Internet community. The main goal of the initiative is to reach consensus on the steering committee for the IPDR prototype definition, and submit the prototype to the appropriate accredited standards body (or bodies) for acceptance. We expect the prototype to be completed and submitted by the end of 1999.

Get involved!

The IPDR group is open to the Internet community at large. For an up to date summary of the work accomplished, work in progress, the IPDR charter and how to participate, go to www.ipdr.org.