IP Billing… A View from Europe

December, in southern France, IIR held their next generation billing conference event. As a precursor to the event, a workshop entitled “An Introduction to IP Billing” was held in conjunction with the first European IPDR Working Group Meeting hosted by Kimber Lewis. I thought it would be useful to take the opportunity to get a consensus from the audience on their views of IP billing. Tony Poulos from Unisys and I co-chaired the workshop, which included presentations from Limor Schweitzer of Xacct Technologies; Liam Maxwell from Portal; Joseph Wee of ntl, which provides fixed telephony, cable, trunk and Internet services to more than 3 million customers in the United Kingdom and Australia; and Andrew Tiller from Geneva Technology.

Armchair Forecasting

These presenters asked about 80 attendees to respond to the following questions by a show of hands. Half of the attendees represented Western Europe, 20 percent represented Scandinavia and 15 percent represented Eastern Europe. The remaining 15 percent were from outside Europe. About 35 percent of the participants represented network operators and service providers, with another 35 percent from software vendors of various types. Consultants, systems integrators and others totaled 15 percent, with another 15 percent representing hardware manufacturers.

1. Who will profit from selling raw telecom capacity in the next 5 years?

a) Long-haul carriers who own fiber
b) Telecom resellers and traders
c) Broadband local loop owners
d) None of the above
There was not much of a consensus on this one, given that the question did not allow for “(e) All of the above.” However, access providers appear to be in the weakest position.

2. Which type of company is in the best position to profit from broadband IP services?

a) Traditional telcos (BT, AT&T, NTT)
b) Companies owning a local loop (cable, ADSL, UMTS)
c) ISPs, ASPs, Web hosters
d) Content owners

This result reveals not quite such a rosy view for telcos and ISPs, although the access providers were not seen as disadvantaged as in Question 1.

3. Can ISPs be profitable?

a) Yes, if they merge with carriers and sell value-added services
b) Yes, if they become portals (selling books, lingerie, etc.)
c) Pure ISPs won’t exist in 3 years
d) There’s plenty of room for 1,500 European ISPs

ISPs are clearly expected to reposition themselves or suffer the consequences.

4. When will fixed-line bandwidth be free to end users?

a) Within the next 2 years
b) Within the next 5 years
c) Sometime after that
d) Never

There was hot debate on this issue. The question was raised whether we were discussing “real time” or “Internet time,” which seems to accelerate with every passing day.

5. Which pricing policy will be most profitable for broadband IP services?

a) Free (e.g., funded by advertising) except for content purchases
b) Flat fee for fixed bandwidth
c) Different usage-based pricing for different applications (e-mail, voice, videoconferencing)
d) All you can eat for a monthly subscription

6. How much of the IP infrastructure should customer self-care systems have access to?

a) Accessing real-time rating and billing information
b) a + provisioning services from an existing service portfolio
c) a + b + modifying service characteristics (QoS, virtual hosting, friends and family)
d) a + b + c + creating new service types

There is a surprising reluctantance to allow customers “deep” access to IP infrastructure, given that service creation has been available within VPNs and in the IN space for a few years, and that “IP to the desktop” may be just around the corner.

7. Should billing functionality be provided by a single vendor?

a) Yes
b) Yes but open architecture and open-standard APIs must evolve (in order to allow competition for specific functionalities)
c) No, we should integrate best-of-breed solutions
d) In-house is the only way to be competitive

8. How important is real-time rating for IP services?

a) Absolutely essential, all the time
b) Quite important for some types of services
c) Over-hyped—near real-time rating is OK
d) What’s real-time rating?

After lengthy discussion on the nature of “real-time” rating and whether “to batch or not to batch,” option (b) won on a recount.

9. How important is real-time billing for IP services?

a) Absolutely essential, all the time
b) Quite important for some types of services
c) Over-hyped—near real-time billing is OK
d) What’s the difference between real-time rating and billing?

Following on from the previous question, clearly the other uses to which rated usage data can be put are more important than “hot billing” per se.

10. What are the key success factors in delivering bundled services?

a) Cross discounting and real-time rating (roaming CDRs come with 1-month delay)

b) Provisioning (across different technologies from a single customer-care system)

c) Creating different complex business offerings using the same basic provisioning and billable event definitions
d) Supporting composite customer types

Bundling issues in Europe are rather different from those in the United States, as the regulatory situations are rather different. Agreement was reached that none of the options above were particularly easy to do well.

11. When would you offer QoS guarantees?

a) Never
b) Only for traffic on my own network
c) Only for traffic carried on my network and networks owned by trusted partners
d) Any time, any place, anywhere

Good to see that ‘trusted partners are actually trusted. However, SLAs and associated penalties between partners trusted or otherwise could certainly be passed on to customers.

12. Why is metering of application-level effective QoS important?

a) Not important—you only need transport layer metering of latency, jitter and throughput
b) Important because the end-user doesn’t care about the underlying telecom infrastructure
c) Interconnect IP traffic settlements can be better addressed at the application level
d) IP QoS is an oxymoron

13. Why is billing for QoS and usage important to next-generation telcos?

a) Not important—flat fee is the way to go.
b) It allows for end-to-end QoS and enables new services based on voice and video over IP
c) It allows the financing of bandwidth capacity growth

d) QoS is either good enough or the customer goes elsewhere

A good debate between (b) and (d). However, it makes you wonder how flexible /scalable their infrastructure really is in supporting these resource intensive/ bandwidth hungry apps.

14. Should usage-data collection be done by the billing system?

a) Billing systems cannot (or should not) handle real-time IP usage events
b) Some billing systems can do it, but ours can’t
c) Yes, for up to a few events/second
d) Yes, modern billing systems can handle any type of input or throughput

Presumably there was a certain amount of bias from the billing vendors present, given the strong response to option (d). Otherwise it was clearly a vote for independent mediation.

15. What are the advantages of using probes to capture usage-data?

a) They can provide the same information as logs from network elements with less overhead
b) Probes provide traffic metrics that network equipment vendors don’t put on their top “to do” list
c) There are no advantages of probes if the logs provided by application servers, firewalls and routers contain the right metrics
d) Probes should never be used

16. What is the most challenging thing about creating IP CDRs?

a) Associating customer IDs to a billable event
b) Condensing 500,000 billable events per second to a manageable number for billing, while making sure that the bill-related details are not lost
c) Collecting from hundreds of distributed information sources with no standard accounting mechanisms
d) There are no challenges left unsolved

The responses to Questions 15 and 16 suggest that the debate on IP mediation and aggregation methodology will continue for a while yet. Moving from network-centric to customer-centric usage data gathering and modeling in a standardized environment will clearly be of benefit to (almost) everyone. But ensuring that time-to-market competitive positioning is preserved will become an issue.

17. Will VoIP replace PSTN?

a) PSTN will remain forever—QoS will never be business-grade
b) Yes, but only in the backbone—terminals/handsets will remain on the existing local loop
c) Within 10 years the PSTN will be gone
d) Video over IP will replace telephony (delivered over ADSL, cable, UMTS)

Apart from the retrograde but rather elegant concept of “switched Internet” to ensure IP QoS, the main focus (or hope) seems to be the development of RSVP and similar methodologies. Author’s Note: I am not sure that I want to have video as my ‘default medium of choice’ for communications. The issue is to provide the right communications processes to suit each situation within which ‘delivery’ occurs.

18. Will Web-based billing replace paper invoicing?

a) No, but it will be an important value-added service
b) No, because the regulators wont allow it
c) Yes, for business customers
d) Yes, all customers will be billed via the Internet

The Road Ahead

With the current divide between U.S. and European telecoms in their perceptions of the value chain, various opportunities and challenges lie ahead.

An obvious one—In the United States the Internet is perceived as a business medium; in Europe it is by and large still seen as an information distribution medium. The European viewpoint is changing rapidly, but these perceptions still dominate the mechanisms and business processes currently capable of extracting sustainable business-to-business revenue from the Internet in Europe. A cultural example of this is that for a majority of Europeans it is still considered almost offensive in a sales situation to suggest that a prospect “go check out our Web site if you want to know about what we do” (“Hey, you’re selling to me. Don’t make my life difficult and make me spend money!”). European business processes typically require the “download/print/get-signature/photocopy/hole-punch/file” sub-routine for anything requiring authorization—something which is often forgotten by Web designers. Business over here has not exactly reached the e-life dream state yet. IP is, of course, a different matter altogether.

A warning—It is something of a simplification, but in general the development of telecom data processing in Europe (particularly for interconnection) is driven by network usage data management, while in the United States the market has been driven by customer (OSS) data manipulation and exchange. There are many reasons for this—compare the Telecommunications Act in the United States with the very different regulatory models, and the historic prevalence of cascade billing and local call charging in Europe (not to mention GSM!)—but currently the divide in perception is greater than it has ever been. Example: in the United States, CCB is seen as a subset of OSS, but in Europe the reverse holds true. Although both markets have much to learn from each other, failure to recognize these differences in perception and practice will severely hinder software vendors who are looking to expand into new markets in either direction across the Atlantic.

An opportunity—When the euro becomes a “usable” pan-European currency in 2002, corporations with multinational presence in Europe are likely to find themselves unable to leverage the real benefits, unless they can come to terms with the integrated processing of multiple tax standards. (Tax harmonization is an issue currently working its way into the European Community political arena, but don’t hold your breath waiting for an outcome.) Assuming they can deal with multiple languages and cultures, U.S. software vendors entering the European market enjoy a remarkable opportunity to exploit their familiarity with 6- and 7-tier tax accounting processes.

A challenge—The adoption of IPv6—which won’t be easy given the need to overlay IPv4 for an indeterminate interim—offers some really exciting possibilities for using IP addresses in ASP and developing microbilling applications. Bear in mind that while the IPv4 address space has not been a major issue in the United States because of the way addresses have been allocated, it is beginning to cripple European and Asian development—there just aren’t enough addresses to go around. IPv6 could fix this (as well as perhaps quite a lot of other things), as it offers 10,000-plus addresses per square meter of the earth’s surface. (But remember what we did to 0800 numbers…)

One for the future—Lastly, and I am not sure if I dreamt this: voice-XML. Think about it—and let me know if you’re working on it! A single document format and output style sheet for user-defined, mixed-media information “display” and retrieval is a really great idea, particularly if it evolves into a bi-directional standard for next-generation UM. Take, for example, the requirement for hands-free, vision-free, in-car WAP phone services. What better way to integrate voice with text, icon, image and video?

Hugh Roberts is consultant conference director for CCB events, IIR Telecoms & Technology London. He can be reached at hugh@hughroberts.com.