Microsoft Messenger Means New Carrier Opportunities

by David S. Isenberg, Principal Prosultant(sm),, inc.
January 18, 2002
Prosultant is a service mark of, inc.

Microsoft's new Messenger client introduces a set of Internet-based communications capabilities that will open ways for telecommunications service providers -- traditional and new -- to earn new revenues and create new value. 

Messenger begins with Internet telephony.  It provides PC-based Internet telephony that is a leap forward in voice quality and ease of use.  Previously, Internet telephony had a toy-like voice quality.  Set-up was quirky and complicated.  Today the impetus of Microsoft's technical advances and marketing muscle is accelerating Internet telephony towards mainstream, everyday calling. 

Microsoft’s Messenger client is an integral part of Windows XP, where it is called Windows Messenger.  It is also available for use with other versions of Windows (and even for Apple's Mac) as part of MSN Messenger.  In the future, analysts expect that Microsoft will release the Messenger client for other environments as well.

Microsoft's Messenger lowers important barriers to PC-based Internet telephony (e.g., latency, difficulty of use).  Windows Messenger voice sounds better than plain old telephony in some cases.  An intuitive user interface and a stone-simple set-up procedure makes Windows Messenger telephony easy to use.  As the Microsoft Messenger client hastens the growth of Internet telephony minutes, it will increase the demand for gateway services between the Internet and the Public Switched Telephone Network (PSTN).

But the importance of Microsoft Messenger goes beyond gateway minutes.  Messenger changes the way we make telephone calls.  Indeed, it could change the way we communicate.  It delivers new functionality and fast-track innovation capabilities.  Carriers that understand Microsoft Messenger will be able to participate in new opportunities in an expanded communications space.  Carriers that ignore or misread these changes are likely to find their old ways of doing business decreasingly profitable. 

In other words, carriers can use Windows and MSN Messenger to generate proven revenues in the VOIP-to-PSTN marketplace.  Then, once the Messenger platform is in place, it will become a launching pad for future services -- and new revenue opportunities -- as they are discovered.

Market Size

Windows Messenger is tightly integrated into every Windows XP system.  XP is expected to become the dominant PC operating system over the next five years, reaching over 120 million computers[1].  Furthermore, industry observers expect Microsoft will extend Windows Messenger to mobile and handheld platforms, set-top boxes, and game consoles. 

Even more importantly, the MSN Messenger client, which is available via download from Microsoft, can be used with most available Windows-based PCs because the client supports Windows 95, 98, Me, 2000 and NT 4 operating systems in addition to Windows XP.  To round out ubiquitous support, there’s even a version of MSN Messenger available for the Apple Mac OS.  This broad support for many operating systems means that most Web-enabled PCs that meet Messenger's (fairly standard) hardware requirements can run a version of the Messenger client. 

Microsoft Messenger users will need to call plain old telephones.  It is likely that users of  Messenger will become addicted to basic features like one-click dialing from their contact list.  Therefore, the first carrier opportunity is to become the gateway between  Microsoft Messenger's Internet telephony and the plain old telephone network, originating and terminating Messenger PC-to-phone minutes.  Within five years, there could be several hundred billion yearly originating and terminating Messenger PC-to-phone minutes[2].  Microsoft opened this market in October 2001 when it announced several carrier partners for terminating Windows Messenger Internet telephony minutes to the PSTN including Net2Phone, Callserve and deltathree.[3]

Future Windows and MSN Messenger services will offer carriers larger opportunities.  Revenue potential for new services could become greater than basic call revenues over the next five years[4].  These include:

·        IP-based Centrex-like services for small businesses,

·        location-based presence services,

·        personal reachability management services,

·        IP-based voice and multimedia conferencing

·        Web co-browsing,

·        Web click-to-talk,

·        Unified messaging

·        Voice portal services

·        etc. -- (to be discovered)

At last October’s Windows XP launch event, Bill Gates chose Verizon to demonstrate how Windows Messenger would change telephony.[5]  Verizon personnel demonstrated how a Windows Messenger interface to Verizon services could be used from a remote location (anywhere on the Internet) to change forwarding status of a user's home phone, gain access to Verizon voice messages, get real-time alerts of incoming calls, and custom-manage home-phone reachability on a caller-by-caller basis. 

Underlying Technology

Session Initiation Protocol (SIP) is the key technology underlying Microsoft Messenger.  SIP is an industry-standard protocol for Internet communications[6].  It defines how client-to-client sessions -- from phone calls to document sharing -- are established across the Internet.  Some of SIP's modular functions include determining presence of a known party, establishing Internet telephone calls, and setting up instant messaging, video conferencing, document sharing, and whiteboard collaboration sessions.  These sessions can be set up individually or as associated streams (e.g., voice plus whiteboard).

SIP is consistent with other Internet-based successes -- it is simple, modular, extensible and scalable.  SIP works through servers that contain Internet addresses (e.g., 222.333.22.33) of the various endpoints in a session.  Since users (e.g., are not always at the same Internet address, a primary function of SIP is to determine where an invited party is on the network when a session starts.  SIP will also play a major role in next-generation wireless systems, since it has been designated the call control standard for 3GPP and 3GPP2 networks.

Once location is established, the SIP device, or a proxy server that represents the device, sends a description of the desired session (voice, whiteboard, etc.) to the invited party's device (or to that device's proxy), which can accept or reject the session.  The session is established when the invited party accepts it.  In the establishment of each SIP session, there is a negotiation to determine the capabilities of the session's endpoint devices.  So, for example, if a telephony session is established between endpoints with both audio and video capabilities, and later in the session the callers decide to add video, then SIP can establish a video connection by issuing a newly specified re-invitation.

The resulting capability -- the ability to reach a person regardless of location or device -- is the realization of a long-held telephone company dream. 

Microsoft's user interface for the Messenger client is based on an instant messaging (IM) window on the user's screen.  IM is becoming a preferred workplace communications medium -- IM minutes in the workplace increased by 110% in 2001[7]

The Microsoft Messenger IM application does not only the usual user-to-user instant messaging, it also mediates SIP session startup and does signaling to control SIP capabilities.  Session set-up is intuitive.  A presence protocol tells the user when his or her contacts are available.  The user that wants to set up a session first determines that the intended session participant is online.  Then the caller sends a special kind of instant message -- an invitation to talk.  This invitation appears to the invited party as a hyperlinked field in the invited party's IM window.  If the invited party clicks to accept the invitation, the session begins.  Or the invited party could reply to the caller via text, e.g., "Try me again in ten minutes." 

If both users are using Windows XP version of Messenger, either party can add additional streams as the session progresses.  One party might offer to show the other a spreadsheet or a website.  Or the two parties might agree to upgrade the call to include video telephony.  In all cases, additional media streams are offered by one party to the other as hyperlinked fields in the Instant Messaging window.  This can be as simple as dragging an open document to the Instant Messaging window and dropping it there.  Or it can be initiated with pull-down menus.  The second party clicks on new hyperlinks as they are offered and the session is redefined to include the corresponding new media.  The experience is of a single, infinitely expansible, continuous interaction.

How Carriers Participate

Carriers that wish to participate in these new SIP offerings that Microsoft offers with Windows and MSN Messenger can begin by provisioning call origination and termination services on SIP.  Then, once the SIP infrastructure is in place, carriers can use the rapid application development capabilities of SIP to expand their base of revenue generating network services.

Interworking between SIP and the PSTN is well understood.  Several vendors, including Cisco, Sonus and Nuera, offer SIP-to-PSTN gateway products.  These come in two basic flavors -- access gateways and trunking gateways.  Access gateways interface SIP to PSTN using the ISDN D-channel.  This provides an entry-level product that can be engineered flexibly for smaller offices and/or to meet local requirements.  Trunking gateways, on the other hand, support higher port densities; they interface directly to legacy SS7 Signal Transfer Points via native A-links. 

Both kinds of gateways support all features needed to extend and complete legacy calls across the Internet-PSTN boundary, such as packet-to-circuit voice transcoding, translation of telephone numbers to Internet-style addresses, generation of in-band tones and announcements, generation and interpretation of ISUP messages, providing voice cut-through at the appropriate point in the call, and handling the various non-completion conditions. 

dynamicsoft's Carrier Solution

Unlike the PSTN, which is owned by known entities operating standardized, well-specified technologies, the Internet is a diversely owned network of different kinds of networks.  This means that the provider of even the most basic application-layer services can assume that these services will traverse several physical networks owned and operated by multiple facility providers.  A SIP Edge Proxy (a device that mediates between end-user terminals and SIP service providers) will usually connect to more than one wholesale Voice over Internet provider.  This creates routing and accounting issues.  Furthermore, security is a critical concern because the SIP infrastructure is exposed to the public Internet, where, in principle, it is accessible by every Internet user. 

dynamicsoft introduced the idea of carrier-class SIP networks in 1998, when the company was founded with the goal of creating a new a communications infrastructure that embodied and extended the Internet's strengths.  dynamicsoft's early carrier experience with SIP networks has shown that carrier requirements extend beyond the basic SIP specification in three areas: routing, accounting and security. 

Routing: Under SIP, sessions are routed based on the location of the invited party and the Internet address of the invited party's proxy.  When call volumes grow, route set-up can be slowed by network elements that are busy or unavailable.  Other kinds of routing issues affect operating costs.  For example, when a carrier operates multiple facilities for redundancy, or peers with multiple service providers, load balancing is desirable, but a flexible load-balancing scheme that favors low cost facilities is even better.  The logic of low-cost load balancing can be extended to redundant network elements of all kinds.

dynamicsoft has built a carrier-class Route Engine that can set up as many as 3,000,000 sessions an hour.  It uses awareness of the application, the characteristics of individual sessions, and the parameters of the carrier's network to make optimized decisions that drive up network throughput and drive down costs.  When a route is discovered, it is cached so future sessions can be established efficiently.  Routes can be based on SIP addresses or on PSTN numbers -- this allows communication between SIP and non-SIP endpoints and the re-creation of classic telco services, e.g., 800 dialing, within the SIP context.

Accounting:  The Internet creates a separation between facilities and services.  In so doing, it introduces the liklihood of third party communications applications -- indeed it makes third parties virtually impossible to exclude.  Carriers need to know (a) how to bill their customers, (b) how to allocate costs among providers of network facilities and service peers, and (c) what kinds of advanced services were used and their billing parameters.  There are additional issues specific to Internet telephony.  For example, customer-use minutes are not tightly tied to network facility parameters.  (For example, a minute of speech could use anywhere from 5 to 64 kbit/s of facility bandwidth.) 

dynamicsoft has created a carrier-class billing mediation methodology to serve the needs of the wholesale Internet telephony provider.  Under this methodology, a special SIP Session Event Record is created at the SIP Edge Proxy.  This Session Event Record logs the unique session identifier, and the start and end times of the call.  It also collects other session-specific information, for example, it uses the persistent Internet address of the service provider to record which provider's facilities carried the call.  The Session Event Record is extensible to track additional customer variables, network resource usage, and usage of additional application resources.  It is combined with the Call Detail Record (CDR) of the switch to create a SIP Wholesale CDR that meets the requirements of both new and legacy billing systems.

Security:  The Internet's separation of services, facilities and customers creates security issues that don't exist in the PSTN.  Normally, Internet service providers and companies use firewalls, which allow certain Internet addresses to pass through them. In many cases this allows too much access, especially where real-time media streams must be allowed.  Hackers can embed information within streams to spoof the system to gain access to subscriber authentication, service authorization, service provisioning, accounting and billing information.  Once accessed, this information can be stolen outright, or used to steal services.  Or the information could be maliciously corrupted to attack and potentially disable services. 

There is a second reason why Internet security is a critical concern.  Network elements facing the Internet are targets for denial-of-service (DOS) attacks that could render these elements dysfunctional or inoperative. 

dynamicsoft has created a  carrier-class SIP security solution.  The solution features enhancements to its Edge Proxy, a new dynamicsoft Firewall Control Proxy and a new kind of firewall.  The new firewall permits traffic passing through it to access only one network element -- the SIP Edge Proxy.  When the Edge Proxy receives a SIP invitation, it authenticates and authorizes it.  Then the Edge Proxy passes it to the Firewall Control Proxy (FCP), which examines the SIP header to determine that the requested application is permitted (for example, if it is a voice call invite).  If so, the Firewall Control Proxy tells the special firewall to create a media "pinhole" so e.g., the voice stream can pass through it. 

There is even more to it.  For example, the Edge Proxy masks critical fields in outgoing headers to prevent unauthorized discovery of inside-the-firewall network topology.  And because firewall control is so critical, the FCP is always hidden from the outside world, accessible only through the firewall and the edge proxy.  In the call set-up process, the firewall and the FCP communicate via special protocol designed to add another layer of security to firewall control.  The result is carrier-class security within the public Internet.

Pay as you grow.  All three functions -- routing, accounting, and security -- work together to provide linear scalability with automatic fail-over and automatic route discovery managed by dynamicsoft's load balancing technology, so carriers can implement SIP infrastructure growth without single points of failure.  Return on investment for an entry-level (but fully carrier-class) SIP network can be measured in months.  dynamicsoft ensures continuous network growth that follows the growth of revenue.  The SIP infrastructure ensures rapid participation in new service opportunities as they arise.

dynamicsoft's  Microsoft Messenger Solution

Until now, this discussion has focused on SIP network requirements for wholesale carriers and large enterprise networks.   But the demands of serving Microsoft Windows and MSN Messenger SIP terminals require that participating carriers must extend SIP functionality towards the end-user customer. 

Consider, for example, the prepaid calling application, in which a caller originates a SIP call from a Microsoft Messenger PC out to a plain old telephone on the PSTN.  Not only must the call be routed from the Internet to the correct PSTN number via a carrier chosen according to user-defined (or other criteria, e.g., least-cost routing), but the caller must be securely authorized to verify available funds and examine billing information in real time.  There must be third-party call control to insert announcements (as funds are depleted or other user-specific or call-specific conditions occur) and terminate the call should funds run out.  Furthermore, self-provisioning via a Web interface reduces carrier costs; users should be able to set up their own account, change their own account information, add funds, terminate service, et cetera. 

The dynamicsoft AppEngine is the platform on which the main aspects of dynamicsoft's Microsoft Messenger Solution are implemented.  The AppEngine is an open, programmable development platform that uses a Web-like Application Program Interface (API) designed for rapid SIP service creation and deployment.  Logically, the AppEngine sits close to the user, right next to the Edge Proxy. 

dynamicsoft offers a basic application for terminating Windows or MSN Messenger minutes as a turnkey application and as a framework for other, more advanced carrier-class applications.  The termination application does subscriber authentication, authorization, call set-up, call control and call supervision.  It interfaces to an industry-standard billing server for account balance verification and maximum call duration computation.  It provides optional real-time account information to the end-user (via Web or IM). 

The dynamicsoft AppEngine's API makes the termination application easily expansible.  Service developers can upgrade and improve classic telecom services, create new services or add communications capabilities to existing PC or Internet applications.  The API includes support for operations, administration, maintenance and provisioning of all applications written for the AppEngine.

Business Model Growth

Jonathan Rosenberg, dynamicsoft's chief scientist, says, "The killer communications apps of the future have not been discovered yet."  He has devoted his career to the proposition that SIP provides the flexibility to discover and deploy these massively useful future applications. 

The "Death of Distance"[8], brought by radical improvements in network technologies, and the Internet have weakened the foundations of the old telecommunications business model.  Telephone companies around the world are struggling.  New services, even new business models, are needed.  Carriers will need to evolve into the new environment.  The carrier-class SIP platform offered by dynamicsoft allows carriers to begin the evolution process by following known revenue sources, beginning with Microsoft Windows and MSN Messenger minutes.  At the same time, deploying the dynamicsoft SIP infrastructure will prepare these carriers for the undiscovered future.

For more information a Service Provider Solution Brief is available at  In addition, deeper technical detail is available in, "Generating New Revenues with a SIP-based Network Solution for Microsoft Messenger Termination" by Matt Lazaro, solutions manager at dynamicsoft; it can be requested at  For further information, visit or write to

David Isenberg is Principal Prosultant(SM) of, inc., an independent communications analysis firm.  He is author of the influential essay “The Rise of The Stupid Network” on the advantages of the Internet’s distributed intelligence architecture.  Prosultant is a service mark of, inc.

This paper was written at the request of dynamicsoft to support dynamicsoft Messenger Solutions.

[1] IDC, 2001.

[2] This estimate is based on Windows XP growth and Windows Messenger usage, per IDC and dynamicsoft research, which projects three billion PC-to-Phone minutes per month by 2005, with a 70% CAGR, or 60 billion minutes a year in 2006.  We believe that this is a very conservative figure.  Given Microsoft's probable extension of Windows Messenger to handheld, mobile, set-top and gaming devices, there could well be 200 to 500 billion Windows Messenger-to-PSTN minutes per year by 2006.

[4] Dataquest, March 2000, estimates total U.S. 2005 call revenues at $55B and value-added network revenues at an additional $75B.

[6] SIP is defined in IETF RFC 2543, by M. Handley, H. Schulzrinne, E. Schooler and J. Rosenberg, March, 1999,

[7] According to Jupiter Communications, the total minutes U.S. workers spent using the top three instant-messaging applications--from America Online, MSN, and Yahoo--increased 110%, from 2.3 billion minutes in September 2000 to 4.9 billion in September 2001. The number of unique users of instant-messaging applications in the workplace also jumped 34%, from 10 million in September 2000 to 13.4 million in September 2001.

[8] The Death of Distance, Francis Cairncross, Harvard Business School Press, Boston, 1997.

Last Modified 18 January 2002