Authentication, Authorization and Accounting

Overview

Being able to connect any computing device to any other irrespective of location is a blessing coupled with security pitfalls. For an individual, free access to any resources on the Internet is enormously powerful. For anyone concerned with corporate security it is a major headache, exposing the trove of corporate confidential information assets, both to unauthorized access and potential misuse, as well as exposure to viruses, works, and other digital pollution.

The Internet has recently seen the proliferation of mobility. In the past because of the high preponderance of desktop computers, only a few pioneers were exploring the possibilities and challenges associated with allowing users to be connected to (and from) different places in the network as they move with their terminals. The spreading market acceptance of portable computers and mobile wireless devices have made it possible to be truly location-independent and extend the network to the home, hotel, remote office, construction site, vehicle, or any other place outside of the centralized LAN environment. The fact that the users can now be connected to, and roam dynamically between, different points in the network breaks one of the most important assumptions in the Internet protocol, which is that the IP address and the physical location of the user are always the same.

New Requirements for Security

One of the principal obstructions which will arise for these mobile devices is already evident in today's implementation of most corporate security systems-that of access control. Anyone who has travelled with their PC to a sister office and has attempted to link back to their home domain has likely experienced difficulties in accessing their "home" resources.

If a remote LAN is connected to the Internet it is almost certainly protected by a firewall (this is a device which limits the access of unknown visitor devices to LAN-based resources). The identification for access control and authority is normally the IP address of the calling device. Thus, the firewall for a corporate LAN is normally configured to only permit external access by devices which have an IP address which is within a predefined range, or has been uniquely registered. Where the device is one which is normally used within the LAN, and is calling in via some form of Remote LAN Access (RLA) technology, the firewall will identify it correctly and thus grant access to the device. If the same device attempts to establish a connection across the public Internet via an ISP, then it will normally inherit a new IP address-allocated by the ISP for the duration of the session-and thus look like a foreign and potentially unwelcome visitor.

The situation is similar upon arrival of the mobile device at a remote site. In this instance the device is allocated a new IP identity within the subnet in order for it to be accepted by, and receive network services from, the local routers and devices within the subnet. If the device were to attempt to use its "primary" or "home" IP address, these would not be available without prior specific configuration of the firewall, neither would access out via the public Internet.

This situation is further complicated by mobile devices which will exist within an IP subnet only for as long as they are in transmission range of the subnet. Firewalls therefore need to become a good deal more sophisticated in their recognition of mobile devices, if the user is to obtain a consistent access and service experience in accordance with their AAA (Authentication (who am I), Authorization (what am I allowed to do), and Accounting (how do you charge me for it)) profile, regardless of their location.

There is much work currently underway in the Internet Engineering Task Force (IETF) to define the AAA requirements of Mobile IP. An IETF working group was formulated to work on a solution to this problem. Its main goals in the interest of a workable solution are:

Convergence of Two Industries

Many in the telecommunications industry see the IETF work as a duplication of effort. Cellular network operators have been able to authorize, authenticate, and account (bill) for voice call roaming for nearly a decade, and GSM data services, such as SMS, are already enabled for roaming across national boundaries with all the inter-network and customer billing taken care of within the infrastructure of the network.

In the telecommunications world, standards are enforced by regulation; however, in the Internet world, standards exist through consensus implementation. As mentioned previously in the Content Provision section, MWIF has been established by industry players to bridge the old world of billing and customer service across into the new, IP-based world.

This leads into the second major consideration for support of mobile users, which is the provision of the new multimedia applications containing voice and video over IP. This "revolution" has been created because of the high speed of processors, together with the promise of UMTS and other 3G standards, enabling the construction of very sophisticated desktop, laptop, and handheld mobile devices, together with the availability of networked multimedia applications. However, the Internet was originally designed and built to handle data flows exclusively in a "best effort way", with no warranties of minimum throughput and delay end-to-end. For video and voice applications to work adequately, the network needs to provide a minimum bandwidth and delay along the path to transmit those flows correctly.

The networks have recognized and anticipated the growth of the mobile data revenues they will soon be able to make, but this is only on the transportation of the data, which compared to the value of the data and transactions they will be carrying, is small beer indeed. Analyst estimates of the annual value of consumer goods and services transacted over mobile networks by 2003 vary between $8billion and $13 billion.

However, there are many barriers to m-commerce happening. The global reach of the technology hits local legal and cultural hurdles which make worldwide standards in commercial rules and best practices difficult to achieve. Telecommunications companies are familiar with doing business globally and as a result all the Wireless Network Operators are keen to raise their role above that of the "bit carrier" and participate centrally in the provision and completion of all of their customers' m-commerce transactions.

Telecommunication billing systems are already equipped to carry out most of the AAA work required for routing traffic and completing a secure transaction. As a result, mobile network operators could soon be able to provide a whole new range of payment schemes for goods bought via mobile devices hosted on their network. However, this assumes the merchant at the other end of the transaction is prepared to use the network's billing system.

And here is the problem: Who “owns” the customer?

Up until now it has been the network operator which “owned” the relationship with the mobile device user through the air-time contract. Subscriber identity is associated with the device in this model. But once IP is delivered to the handset (GPRS should deliver this within 2000) then the network operator's role within m-commerce can be likened to that of the average ISP, purely providing the transport for a nominal subscription fee and charge per packet. An end-to-end m-commerce transaction can take place without the network being aware of it.

Mobile network operators are thus faced with two main problems: how to provide the full range of AAA services for IP-based m-commerce, and how to work with vendors and service providers to offer the AAA capabilities of security and micro-payment while not damaging their relationship with banks and credit card companies which are providing these services for the fixed-connection Internet.

Another example of convergence can be seen in the groundwork currently underway for the roll-out of GPRS and EDGE, packet-based 2.5 generation global cellular standards. Wireless and cellular networks are moving to a purer IP-based world, thereby laying the foundation for more open standards and greater interoperability. Early in 2000 a group of service providers and suppliers formed the Mobile Wireless Internet Forum (MWIF), a forum focused on accelerating the deployment of open, Internet-based standards for mobile wireless networks.

MWIF aims to provide a forum to identify and resolve issues surrounding the development of key specifications based on IP wireless networks, independent of air interface, to enable seamless integration between mobile telephony and other IP-based services such as data, voice, video, and multimedia. The first goal is to produce a reference architecture that supports open interfaces across access technologies through a common IP core, and a distributed, peer-level scalable IP-based architecture that includes IP standard interfaces for billing and customer service.

Specifications developed by MWIF will be available to traditional standards bodies and other industry groups. MWIF also works to ensure that key regional and global standards groups address requirements that operators globally see as critical to early and economic deployment of IP mobile wireless networks. This focus forms a basis for the issues of content distribution from the traditional Internet and the wireless and mobile world.

Requirements for Secure Transactions

There are four different concerns that a security system can address: privacy, integrity, authentication, and non-repudiation.

Privacy: Ensures that only the sender and the intended recipient of an encrypted message can read the contents of that message.

Integrity: Ensures the detection of any change in the content of a message between the time it is sent and the time it is received.

Authentication: Ensures that all parties in a communication are who they claim to be.

Non-repudiation: Provides a method to guarantee that a party to a transaction cannot falsely claim that they did not participate in that transaction.

Whatever security is provided for mobile commerce, there needs to be some level of positive interaction between the user and the device. Although an offline PIN can be stored, it must be the user that initiates its submission.

Over the Internet, the Secure Socket Layer (SSL) protocol, digital certificates, and either user name/password pairs or digital signatures are used together to provide all four types of security.

Digital signatures are indispensable for making electronic communications legally binding. In comparison with other authentication techniques, those using one-time only passwords (PIN process) have proven themselves extremely cumbersome. Furthermore, they do not offer any real security for the transaction content because there is no association between the PIN and the message.

The need for an alternative model for mobile phone usage is best understood by revisiting the differences between the fixed connection and wireless Internet environments. The fixed connection world assumes that the client device has robust processing capabilities and the communications interface is always on, always connected, and has at its disposal high bandwidth at low latency. The very nature of wireless communications is almost the reverse where bandwidth is low, costly, has high latency, and is unpredictable and of variable quality, even to the point call breakdown.

As with many aspects of the deployment of mobile and wireless data communications, the architecture for which SSL was designed is not sustainable in the wireless environment. Mobile phone users would be disappointed by the delays required to process SSL transactions, and building in to handsets the processing power needed to support SSL capability would raise the unit cost of devices which rely on an affordable pricing model to accelerate market growth. So, alternative mobile-optimized methods are required.

With security requirements being so much greater for hand-held and wireless devices than for their desktop or fixed connection counterparts, the use of the GSM architecture for digital signatures has significant benefits. Each mobile telephone has its own smartcard or SIM card which identifies the user for service and billing, provides a storage medium for the private keys, and, as the key never leaves the card, it meets the prerequisites for a high degree of security. Current implementations of TDMA and CDMA have no accommodation for equivalent facilities. In discussion with a leading provider in the CDMA space, we learned that they are wrestling with how to provide similar functions, or indeed interoperability, with this useful consumer construct. The existing SIM manufacturers are working hard to provide continuity for SAT in the 3G standards-this may well be the opportunity to extend these benefits to the TDMA and CDMA world.

Taxonomy

WAP Security

In June of 1999, the WAP Forum formally approved WAP Version 1.1. which includes the Wireless Transport Layer Security (WTLS) specification. WTLS defines how Internet security is extended to the wireless Internet, and is intended to give network operators, application developers, and vendors a way of building the confidence of end-users and thus open whole new markets to e-commerce in the same way as the arrival of SSL did for the fixed-connection Internet.

The WAP security model and the Wireless Transport Layer Security (WTLS) mechanism are based on the Internet standard security protocol TLS 1.0, which in turn is based on SSL 3.0.

WTLS was specifically designed to conduct secure transactions without requiring desktop levels of processing power and memory in the handset. WTLS processes security algorithms faster by minimizing protocol overhead and enables more data compression than traditional SSL solutions. As a result, WTLS can perform security well within the constraints of a wireless network. These optimizations mean that smaller, portable consumer devices can now communicate securely over the Internet.

A secure WAP conversation occurs in two stages. First the transmission between the web server and the WAP gateway occurs over SSL. The onward transmission of this message over the air interface to and from the WAP browser device is over wireless networks using WTLS. Essentially the WAP gateway serves as a bridge between the WTLS and SSL security protocols.

There are considerations (some say limitations) in the introduction of a bridging protocol like WTLS. The current WAP security model requires a strong relationship between the network operator and the content provider. The WAP Forum has recognized that as the market for highly secure applications increases, a more flexible and extensible solution will be needed. When working across many different wireless networks, application developers must be assured that their content remains encrypted from the time it leaves their application server until it arrives at the WAP handset. As a result there is a process underway to develop this more advanced security solution, which must address the enterprise's need for higher security and the operator's need for proper integration with WAP gateways in the wireless network.

The WAP Forum is essentially a telephony standards committee, which is operating in an Internet world where standards are adopted by implementation rather than regulation and certification. There is therefore some consternation that solutions are coming to market in advance of the WAP Forum having established a standard approach to providing end-to-end secure content. WAP gateways are designed for use in an operator's network; however, many of these new approaches promote installing a WAP gateway at a content provider or in an enterprise. As well as potentially diluting the network operator’s hold on the customer-a development which many operators oppose-this creates a number of interoperability issues for content providers, subscribers, and wireless network operators. The WAP Forum members have recognized the problem and are seeking to address it.

Already, key Internet applications for handsets have been deployed, including e-banking, stock trading, e-commerce, and other exchanges of private and mission-critical data.

With WTLS optimized to ensure transactions are conducted in a secure and user-friendly way, and WAP-capable handsets reaching the market, subscribers are beginning to embrace wireless e-commerce in the same fashion that consumers adopted wired e-commerce over the last 18 months.

SIM Application Toolkit

The SIM Application Toolkit (SAT) is a GSM feature which was integrated into the GSM standards in Release 96, with further enhancements added as part of the Release 97 feature set.

The SIM is the Subscriber Identity Module, a computer chip on a plastic card permanently inserted into a GSM mobile phone containing a profile of the subscriber for billing, and so on. The SIM Application Toolkit defines a set of commands which enables an application on the SIM card to communicate with the mobile phone in a standardized format. Furthermore, cryptographic functions can be installed on the SIM card, enabling certain messages to be encrypted or signed.

The commercial driver behind the creation of SIM Application Toolkit was the requirement for network operators to be able to offer discrete value-added services, without the need for handset manufacturers to build network-specific models. This is possible because it is the network operator that owns the SIM. The networks requested that the SIM architecture be extended to allow for the placement of applications in the memory, and most importantly, that these new SIMs could work on any Mobile that supports the Toolkit feature.

Logically, the SIM Application Toolkit is completely separate from the GSM functionality on the SIM. There is a distinct set of commands between the Mobile and the SIM specifically for the toolkit. This allows the SIM toolkit on the handset to communicate in a client/server model independently of the GSM telephony communication. This logical split allows and encourages third-party applications, as well as operator-specific applications, to be resident on the SIM, with a co-located application part in the network or even outside of the network. Further, the toolkit can interact directly with the mobile itself, by adding to the mobile's menu structure. Other examples of handset behavior which can be controlled using SAT include sending a short message; setting up a voice call to a number held by the SIM; setting up a data call to a number and bearer capabilities held by the SIM; sending a Supplementary Services (SS) control or USSD (Unstructured Supplementary Services Data) string; playing a tone in the mobile's earpiece or ringer; initiating a dialog with the user. An optional "help" feature is also available so that additional help text can be displayed to the user if requested.

Local information, about cell identity, call status, coverage status, and so on, can all be used as the basis to provide value-added services to the subscriber, and can also be fed to the SIM application as and when the current local situation of the handset changes. The network can then provide an application to modify the handset behavior accordingly.

Communication between the toolkit application on the SIM and its "other half" on a server somewhere in the network is currently achieved by using the Short Message Service (SMS). This combination of SMS and the SIM Application Toolkit promotes the development of mobile phone applications which considerably ease the process of entering data (for example, a menu-driven bank system). In future, SAT will use other transport mechanisms, such as Unstructured Supplementary Services Data (USSD) strings or the General Packet Radio Service (GPRS).

Until now, each card needed its own development environment; however, the client/server implementation model of SAT also means that Java SIM cards can be used. With the JavaPhone API, developers can develop telephony features that are crucial to integrating value-added computing capabilities without sacrificing the instantaneous telephone service consumers expect. The Java Phone API will include power management, display controls, and data synchronization. The use of common Java applets has enabled the provision of compression algorithms, encryption/decryption, PKI interfaces, and the capability to validate digital signatures. which can be used to develop secure SAT applications for mobile phones.

Applications stored as Java applets will further encourage developers to produce applications, as they will enjoy the benefits of Java technology (namely operating system-independence) when programming SIM card applications. Applications only need be developed once and can then be loaded onto any card. Using SMS as a bearer, menus of existing applications can be updated or new applications downloaded onto the card. This approach reduces time-to-market and development costs.

Using the above, it is now possible for the SIM toolkit feature set to include data download to the SIM from the network, where an SMS is delivered directly to the SIM from the network, which could contain data or commands distinct to the toolkit application, or, indeed, allow a new application to be loaded into the SIM.

Considering the scale and variety of SIM-resident applications that will be available in the future (banking services, airline booking and inquiry services, and so on), some of which are commercially or security sensitive, the GSM SAT specification has been developed which standardizes the communication between the end-to-end entities in a secure way for SMS, so that it will be impossible for any messaging between the toolkit application and its counterpart in or behind the network to be intercepted, read, or modified.

Industry Groups Focusing on the AAA for Wireless and Mobile

In addition, to the efforts of the IETF, a number of other organizations are active in defining the needs of the security area.

PKI Forum

Public Key Encryption is seen as a critical component of the emerging Internet security and trust infrastructure. PKI (Public Key Infrastructure) enables users of an unsecured public network such as the Internet to securely and privately exchange sensitive data and conduct financial transactions through the use of a public and a private cryptographic key pair that is obtained and shared through a trusted authority. The Public Key Infrastructure provides for digital certificates that can identify individuals or organizations and directory services that can store and, when necessary, revoke them.

The Public Key Infrastructure assumes the use of public key cryptography, which is the most common method on the Internet for authenticating a message sender or encrypting and decrypting a message. Traditional cryptography has usually involved the creation and sharing of a secret key for the encryption and decryption of messages.

Use of asymmetric encryption methods requires reliable distribution of public keys. All users must be able to be sure that public keys really belong to those parties they are said to belong to. The solution is Public Key Infrastructure (PKI), certification authorities, and reliable third parties, which offer services. The objective of the infrastructure is to produce reliable security service for the net.

The PKI Forum, founded by Baltimore Technologies, Entrust Technologies, IBM, Microsoft, and RSA Security is an international, not-for-profit, multi-vendor alliance whose purpose is to accelerate the adoption and use of PKI (Public-Key Infrastructure) products and services. The PKI Forum advocates industry co-operation and market awareness to enable organizations to understand and exploit the value of PKI in their e-business applications.

The major objectives of the PKI Forum are:

The PKI Forum operates as an autonomous, unincorporated entity within The Open Group, which provides legal and management services. The PKI Forum aims to provide a forum for vendors to demonstrate support for standards-based, interoperable PKI solutions for e-business. In a model becoming common in the high tech industry, the Forum serves to bring customers and vendors together in a vendor-neutral setting to increase customer knowledge about the value of PKI and demonstrate how PKI solves the security issues for e-business. It is hoped that that Forum will foster interoperability by interacting with appropriate standards and testing bodies and initiating studies and demonstration projects in order to show the value of interoperable PKI.

Through these efforts the Forum predicts it will accelerate the deployment of PKI and PKI-based solutions and communicate the compelling value of PKI as a trusted base for e-business applications.

To accomplish its objectives, the PKI Forum has established various working groups with active membership participation. Special interest groups may also be formed from time to time to resolve specific issues. Given the PKI Forum's initial objectives, at least two standing working groups will be formed. The Business Working Group will identify business-based interoperability requirements. The Technical Group will develop PKI interoperability profiles and champion product interoperability demonstrations to ensure that products are certified against those profiles.

The PKI Forum is not a standards body. Current formal standards organizations that address PKI (such as the IETF and ANSI) and other organizations which have a PKI industry focus (such as ANX, NACHA, BITS, and Identrus) play an important role in developing open standards and in fostering the emergence of sector-based PKI. However, baseline product interoperability is not their primary charter. With its emphasis on building an assured foundation of interoperable PKI products, the PKI Forum will directly support (and influence) the work of technical standards and interoperability profiles as specified by the IETF, The Open Group, ANSI, governmental Common Criteria programs, and industry-specific programs.

Radicchio

Radicchio is an organization founded by Sonera SmartTrust with GemPlus and EDS in September 1999 to promote PKI in secure wireless e-commerce and is well on its way to becoming the industry voice and authority in this space. Radicchio is registered in the United Kingdom, for the benefit of its members which include certification authorities, mobile operators, systems integrators, device manufacturers, and software companies. Radicchio aims to raise awareness of the opportunities presented by secure wireless e-commerce through the expansion of its membership.

The initiative promotes the use of an environment based on a Public Key Infrastructure (PKI), allowing secure electronic transactions to take place over wireless networks. As a consortium of the industry’s leading wireless e-commerce companies, Radicchio will persuade international regulators and government bodies to accelerate the adoption of legislation which supports secure wireless e-commerce globally.

The stated goals of Radicchio are:

Mobile Electronic Signature Consortium

The Mobile Electronic Signature Consortium was formed in January 2000 and is an association of companies and organizations from the mobile phone and Internet sectors. The basic pretext for forming the group is that the founder members assume that the current separation of mobile telecommunications and Internet as implemented in WAP will not last and that the isolated use of various electronic distribution channels will no longer be able to cope with consumer demands for integrated, mobile offerings and services. Only by using the legally effective digital signature it will be possible to perform business via electronic channels in secure and understandable form. The members are all working on the integration of mobile telecommunications and fixed connection Internet technologies to generate services that will require a mobile digital signature as a way to establish legal security for transactions performed.

From the end-user perspective it is intended to add a “signature” button to the mobile phone keyboard. This capability will be a benefit in the marketing of such devices The message this should create in the mind of the end-customer is that only mobile end-devices with a sign button provide access to secure electronic/mobile commerce.

Smartcards possess the storage capacity for the private key. This means that the implementation of a signature application on the card can be achieved by using the SIM Application Toolkit. Alternatively it will create the opportunity for a new card generation. The introduction of new add-on value services in co-operation with service and content providers is thought to provide a significant increase in value for network operators.

The main objective of the consortium is to develop a secure cross-application infrastructure for the deployment of mobile digital signatures. To achieve this objective, key players of the mobile telecommunications and Internet sectors will define a standard that facilitates the integration of the mobile phone and of other mobile devices into the current online world.

Current members of the consortium include BROKAT Infosystems AG, Bank of Tokyo-Mitsubishi, cryptovision gmbh, Deutsche Telekom/T-TeleSec, D-Trust GmbH, E-Plus Mobilfunk GmbH, Gemplus, HypoVereinsbank AG, Mannesmann Mobilfunk GmbH, ORGA Kartensysteme Gmbh, Schlumberger, Siemens AG, Sonera SmartTrust GmbH, TC TrustCenter GmbH, and VIAG Interkom GmbH & Co.

In principle, the Mobile Electronic Signature Consortium is planned as an open forum. Work on the development of respective business rules, among other things, for integrating new members is currently in progress.

The Mobile Electronic Signature Consortium will support application providers that require a legally effective statement of the customer to perform their business. Among these are providers of bespoke, critical services with a high level of security; for example, banks, insurance companies, lotteries, as well as trading agencies for upscale products. The standard to be developed will be freely available so that it can be integrated into already existing applications or be considered in the development of new and innovative applications.

Conclusion and Role of The Open Group

The Open Group should continue to monitor closely the activities of the IETF Mobile IP and Security working group and the Mobile Wireless Internet Forum (MWIF) to assess the extent to which the problems of AAA in the wireless space are being effectively addressed.

Existing Open Group enterprise integration activities addressing Security and eCommerce (LDAP-based), Directory Interoperability, and Enterprise Management are clearly synergistic with the needs of the wireless environment. The Open Group is able to bring together experts from the Mobile Management Forum and its existing programs to ensure that the wireless needs are addressed in a way that ensures smooth integration with existing enterprise systems.