Networks of the future will be more about users rather than a particular network infrastructure, allowing subscribers to link with any available service regardless of what device or access method they are using. Already, fixed/wireless convergence is a growing trend, and service providers operating intelligent IP networks are seeking a path to deliver seamless access to their applications and content, regardless of the subscriber’s location. To achieve this seamless operation, carriers require a pivotal control and management point in a converged network that is independent of network access method (CDMA, GPRS, WiMax, DSL, cable, etc.), subscriber device (work LAN, home PC or set-top box, wireless handset), or application (voice, e-mail, VPN, ERP, CRM). In the IP Multimedia Subsystem (IMS) vision, this central control and management point is the Home Subscriber Subsystem (HSS).
Before they can take full advantage of the attractive revenue enhancements and cost savings associated with IMS networks, service providers need to decide how to transition their legacy networks to an IMS framework. To do this well, it is important to understand the evolution of the hardware, software, and operational infrastructure involved in creating robust IMS central control and management. Instead of a wholesale transition to IMS, it will likely be better to take an evolutionary approach, migrating small numbers of services at a time rather than wholesale implementations. This “slow and steady” approach will give service providers the opportunity to test out new IMS services and realize new revenue streams while still maintaining the functionality of their legacy network.
Changes Required for IMS
By its very nature, IMS requires a rethinking of the network topology. Its wide-reaching scale necessitates a move from a service-level view to a subscriber-centric view in the network, because making IP applications available over all networks requires a focus on the individual subscriber, rather than on the capabilities of any given network.
Implementing subscriber-centric policy management in this multi-network environment involves the management of millions of individual subscriber profiles, possibly hundreds of service and application profiles, and support for tens of thousands of simultaneous user authentication, service access, and mediation transactions. Subscriber-centric operations require a single view of the data relating to a subscriber and specific mechanisms to carry out per-subscriber service policies — including roaming permissions, quality of service (QoS) settings, time-of-day permissions, and data download thresholds — across multiple networks that are owned and not owned by the provider. It is impossible to perform these tasks with existing wireless and wireline legacy equipment, such as an Authentication, Authorization and Accounting (AAA) and/or home location register (HLR). The choice then becomes to upgrade this equipment with basic HSS capabilities or to implement a new robust HSS to create an IP service layer that operates alongside the legacy equipment but operates independently of the underlying networks.
By its very nature, IMS requires a rethinking of the network topology.
AAA systems have traditionally provided simple access to the network. Now, IMS migration provides the opportunity to collapse all of the legacy AAA functions into the HSS, streamlining operations and further reducing expenses. Depending on their feature set, some legacy AAA network elements could be upgraded to serve the IMS architecture. However, these upgrades could still require the network to query multiple databases in order to get a clear view of network activity and involve customized application programming interfaces (APIs) for each application. This type of approach actually works directly against one of the key benefits of IMS — rapid deployment of new services. It also overlooks the opportunities available to service providers when they centralize all of the subscriber data in a single HSS. In the end, most service providers are not likely to do a wholesale transition to IMS, and upgrading the AAA may risk the functionality of existing network services.
Similar to the AAA upgrade, service providers could elect to upgrade their HLR to perform basic HSS functions. In the planning stages of this type of upgrade, key considerations include the overall capacity (in terms of subscribers), the network connectivity (IP or SS7), and understanding the effect of HLR failure on the HSS function, and vice versa. Without addressing these concerns, the upgrade can put the incumbent voice networks and businesses at risk. While it may offer lower costs today, service providers choosing the upgrade path should also consider the long-range financial implications.
Many service providers may find it prudent to maintain their existing HLR and implement
separate HSS functionality for the growing IMS platform, depreciating the HLR equipment and eventually retiring the HLR when the IMS transition is complete.
Most service providers will likely elect to implement a new HSS server for their IMS deployment. This will also be the case for wireline service providers, as they do not use HLRs and it is unlikely that their AAA servers will support wireless devices. Implementing a separate HSS in the network to handle evolving IMS services ensures that the existing network can continue to operate without disruption while new services come online. Over time, as the network evolves more fully to IMS, aging HLRs and AAA servers can be retired.
Before selecting an HSS server, it is important to recognize that they vary in functionality, capability, and operation. For instance, an HSS server running a robust subscriber-centric policy management program is access agnostic, meaning that the same subscriber profile on an HSS server for a cable network can be applied to a WiFi, WiMAX, or mobile access in the future. But not all HSS servers have the capability to support a subscriber-centric network.
For instance, it may be tempting to buy a complete product line from a single vendor. While this may seem like a quick solution, it will almost certainly not be the best one. This is because many of the larger network infrastructure providers are focusing on top-level elements, such as the Call State Control Function (CSCF), and they provide only basic HSS and Policy Decision Functionality (PDF). These types of HSS servers can severely limit a service provider’s capability to bring more complex services online. Instead, the better approach is to use best-of-breed vendors for specific elements, and then ask the major vendor to coordinate with the smaller ones.
Role of the HSS
As the central repository for subscriber and application profiles and policies, the HSS is the make or break point in the success of an IMS network. Enabling call and session enablement, service authorization and authentification, the HSS leverages the information in individual subscriber profiles to enforce policy in the network through a logical chain of processes.
The HSS is the make or break point in the success of an IMS network.
As the subscriber registers with the network, the HSS consults the policy database to learn the roaming permissions, account status, and time-of-day permissions. Based on these, the HSS provides the subscriber with the ability to access the IMS framework. The subscriber then accesses an application, which is authorized by the HSS at the point of request. During the authorization process, the HSS consults the policy database once again to learn the application specific permissions and other policies to be applied to this subscriber. The HSS then authorizes access, directing the call processing elements to allow a session to commence.
This single unified view of the subscriber that a robust HSS element can provide in the IMS architecture translates into significant operating expense savings for the service provider. By acting as the central subscriber database for the network, managing one profile per subscriber, the HSS enables rapid service provisioning and vastly simplifies the addition or deletion of services and subscribers, significantly cutting down on administration time.
For example, since customer data is stored in one centralized database, provisioning time is not dependent on the number of applications effected. The same analysis holds for the case of subscriber deactivation. The savings relating to these operational expenditures can add up quickly, because whenever a new subscriber is added or deleted, the amount of time required for an operator staff member to enter the data is fixed, instead of being dependent on the number of applications. And, in a subscriber-centric system, once subscribers are activated, they can manage their own service profiles online, eliminating the need for in-house staff to make ongoing adjustments to provisioning.
In addition, since everything is now stored in one place, the IT staff only has to manage one single platform rather than multiple platforms, freeing up engineering resources. These savings can be significant, since every database server typically requires one person-day per month of IT support. So if a service provider is running five applications with two databases each, the resulting IT cost would be 10 person-days per month, which can be collapsed into one person-day per month for a single database.
Moving to an all IP infrastructure allows service providers to rapidly deploy new services, and it can give them unprecedented abilities to tailor services for individual customers, connecting subscribers with applications on demand or as subscriber profiles change. This subscriber-centric model uses a policy management system that fine-tunes the network resources in real-time, enforces the service provider’s policy rules, and tracks the resources each subscriber consumes for billing purposes.
Providers can leverage this information to simplify service definition and provisioning and to accelerate the creation of new services and service bundles. New service tiers can be based on anything that is easily measured and simply controlled, such as monthly consumption, bandwidth, and quality of service (QoS). In sophisticated systems, it is possible to enable application-specific tiers, such as tiers with and without voice over IP (VoIP) capabilities.
By building IMS applications that take advantage of the individualized nature of subscriber-centric networks and the robust capabilities of the HSS, service providers can gain a competitive advantage and greatly bolster their ability to retain customers while simultaneously increasing average revenue per user.
Russ Freen is the co-founder and Chief Technology Officer for Bridgewater Systems (news - alert). For more information, please visit the company online at www.bridgewatersystems.com.