From Time and Distance to Bandwidth and Priority

Posted: 11/1999

From Time and Distance to Bandwidth and Priority
By Robert Rosenberg

The financial markets have created a bubble, which seems to envelop nearly every
initial public offering (IPO) of an Internet-related business. Though that bubble will
burst, it nevertheless is apparent the public Internet is the new engine of
telecommunications that initially will augment, and later supplant, the circuit-switched,
public switched telephone network (PSTN). The word Internet conjures up new concepts of
what constitutes a telecom service and a new delivery system for those services, and
creates a foundation for new business development. Using Internet protocol (IP) to build a
telecom service, carriers can, for the first time, establish tiers or grades of service
and create variable pricing to reflect real marketplace conditions.

Insight’s research suggests that at least for the next few years, billing for these new
IP-based services will be an element of a convergent billing operation (see chart below).
Mediation is the process of taking switch outputs and converting them into billable
events. Much like their circuit-switched predecessors, the mediation function for an IP
billing system will format switch output into billable event detail records. The big
difference between the circuit-switched and the IP model is that the IP billing system
must be able to handle many more variables, and these variables must be reflected in the
packet headers.

Graph: IP Billing as a Percent of Total Billing/Customer Care Revenues,
1998-2004

IP-based billing is derived from Internet usage records (IURs), which are similar to
the call detail records (CDRs) of circuit-switched billing. The IUR utilizes fields and
functions from the traditional CDR and extends it to monitor a broader range of parameters
encompassing session length, packet, transaction, service level or central processing unit
(CPU) usage.

The converged bill of today will evolve to a new paradigm of real-time, usage-based
billing based on session length in an all-packet network. The future of IP billing systems
centers on the benefits that flow to carriers and users, some of which are:

* Increased customer satisfaction through the immediate delivery of value-added
information and services. Real-time billing supports the provision of value-added
information services to customers in real time, even if bills are generated only once a
month. Customers can sign up, be registered online, and begin accessing and adding
services immediately. For example, users of online games can purchase additional game
resources while immersed in play.

* Increased revenue for carriers through more accurate and more timely billing cycles
and collections.

* Rapid product deployment. Real-time capabilities allow providers to roll out a
service right away, a crucial capability in today’s markets.

* Enhanced target marketing. An Internet telephony service provider (ITSP) can target
promotions to customers based on calling patterns and can deliver the promotions in real
time, while the customers are online.

* Reduced fraud through real-time detection. Real-time systems protect against fraud by
allowing the network provider to check the user’s available credit while online. These
systems can detect in real time whether a customer’s ID is being used in more than one
place in the network at the same time and, if so, services can be canceled immediately.

Differences of Packet- and Circuit-Switched

There are some fundamental differences between systems designed for Internet telephony
and ones designed for traditional telephony. Traditional telephony operates on
circuit-switched networks while Internet telephony operates on IP packet-switched networks
with IP packets operating over a distributed network. IP services are connectionless,
whereas circuit-switched networks are connection-oriented. Traditional voice services fill
an entire channel or circuit being used. IP packets, however, are scattered across the
network, taking different routes. These packets are lined up at the receiving gateway.
Packet five in a transmission, for example, may arrive before packet three. Packet one may
show up last, and packet seven might not arrive at all and must be retransmitted. The
receiving gateway reassembles messages in the proper order from the arriving packets. As a
result, IP networks are more efficient in terms of network use. Moreover, distance is not
important to IP services, while location and distance is a prime determinant of pricing in
the circuit-switched world.

Another major distinction between circuit-switched and packet-switched billing systems
is the difference in the networks’ technologies and the service provision.
Circuit-switched networks are older, more proprietary, legacy-based systems, while packet
networks are much less complex. From a service standpoint, circuit-switched networks have
a much higher availability (99.9 percent for circuit-switched compared with the 80 percent
availability for packet networks). The quality of service (QoS) for circuit-switched
networks is much higher and the networks more reliable overall than packet switches. In
all IP services, there is a one-to-three-second latency inherent in today’s packet network
technology, which explains why the quality of real-time analog services, such as voice or
video, is degraded in packet networks.

Disparate Quality of Voice Communications

The disparate quality of voice and other connection-oriented communications is the
major difference between packet and circuit-switched networks. The lower quality of IP
telephony services is the reason why IP telephony now is used mainly by students, PC
hobbyists and those with numerous family members living outside of the United States, all
of whom are seeking an inexpensive way to make international calls. The sound cards on
most PCs, if they exist at all, were made for games, not for a voice transmission. Screen
savers, spreadsheet sorting and other background PC-based functions can further degrade
the call quality. In addition, it is awkward to bend over a PC to talk to someone.
Companies do not want to use this technology to speak with their customers. If IP
telephones are easy to set up and do not use a lot of capacity in their data networks,
companies may decide to use them internally. In the United States, however, traditional
telephony prices are so low that saving an extra cent or two per minute may not be worth
the effort of adopting IP telephony.

Flexibility of IP Billing

Billing systems for IP services require much greater flexibility than the systems used
in the traditional circuit-switched billing model. Like their circuit-switched
predecessors, the mediation function for IP billing systems formats switch output into
billable event detail records. The big difference between the circuit-switched and the IP
model is that the IP billing system handles many more variables, and these variables are
reflected in the packet headers.

Insight’s research suggests that most often IP billing systems are an element in a
convergent billing operation–they may be standalone–but more frequently, we expect to
see IP billing systems feeding into one of today’s leading billing and customer care
systems. As a niche-specialty area, IP billing must adapt to more established billing
systems to be successful. Once the records are formatted and translated to a CDR-type of
format, they can be fed into any of today’s leading billing systems for flexible rating
and billing.

The typical IP billing system is likely to run off of client/server architectures.
Unlike most circuit-switched networks, IP billing systems are not wedded to a mainframe
for bill processing. The IP systems in place today tend to be much smaller, with fewer
subscribers than a circuit-switched network. When we do encounter circuit-switched network
providers offering service, they typically segregate IP billing systems from the main
portion of their networks.

IP billing systems give carriers access to object-oriented toolkits and open
application programming interfaces (APIs). This process allows carriers to configure their
own billing systems, resulting in more timely new product introductions and deliveries.
With all the billing system functions performed in house, carriers have greater control
over the programming. They also can eliminate the need for expensive outsourcing and
reduce the danger of revealing proprietary billing system information to external parties.

The newer IP billing systems cost less than traditional billing systems. Still, vendors
of traditional systems such as IBM Corp., Armonk, N.Y.; Hewlett-Packard Co., Palo Alto,
Calif.; and Kenan Systems Corp., Cambridge, Mass., claim the newer IP billing systems do
not have sufficient power to perform the telco applications for larger companies.
Windows-NT proponents, on the other hand, say scalability and redundancy in IP billing
networks are accomplished by adding more processors.

IP billing systems require little, if any, proprietary programming and resources to
support custom development. They are not tied into the legacy systems attending
circuit-switched networks. IP billing systems offer many more parameters that can be coded
into a packet header’s sorting information. Leading billing vendors are offering 20 to 30
different fields in IUR formats, many more than circuit-switched services offer today.

The greatest weakness in IP billing is the difficulty of validating a PC’s IP address.
IP addresses are assigned dynamically within an IP network and do not necessarily
correspond to a telephone number or customer identity. Since many service providers now
are struggling with maintaining number portability, they do not want customers to have
their own static IP address because it is more difficult to maintain. IP services still
face technical problems in fraud prevention and in the complexity of packet-based event
detail records.