August 2007 | Volume 10 / Nuber 8
IMS Interoperability Testing for Solid Roll-outs
By Pierre Lynch
Delivering advanced communications services only works if your networks are speaking to each other.
IMS (IP Multimedia Subsystem) networks will likely become a primary set of next-generation networking specifications for the entire telecommunications industry. Why? The business benefits and opportunities for service providers, their equipment suppliers and partners are extremely compelling. They include convergence of fixed/mobile and voice/data, quicker introduction of new services, lower network maintenance OPEX and the migration to VoIP away from traditional networks.
Telcos will realize more control over their networks, and will be able to implement a scalable, interoperable platform for new value-added services, without watching them being undercut by Internet giveaways. The ability to satisfy demand for fixed/mobile convergence will also drive new business. For instance, IMS-capable networks will allow, among other things, a cellular telephone user to communicate seamlessly using VoIP over an 802.11x access network. This will not only make possible delivery of video-on-demand to corporate desktops over Ethernet and to handheld computers over 802.11 or 3G wireless networks, but also allow blending that service with other services such as presence, unified messaging and location. IMS will let users transparently access their services across cable, wireline PSTN service and wireless service, both locally within a service provider network and roaming between carriers. More good news: IMS is expected to deliver tremendous opportunity for revenue growth for carriers/service providers, both through attracting new customers and by increasing ARPU (Average Revenue Per User). It’s also expected to reduce customer churn and cut out rival poaching on their customer base. This rosy picture is obviously dependent on a successful, rock-solid IMS deployment.
Analysts predict both consumers and the corporate base will be drawn to the promise of IMS, and will remunerate leading-edge service providers with reasonably high-margin revenue. However, if IMS-based services stumble on reliability or disappoint customer experience expectations, repeat business will fizzle. A rushed, untested deployment would likely be unsuccessful due to the numerous protocols in play required to support IMS and the ensuing interoperability challenges, and probably spell disaster for the hastily-minded Telco.
Since service providers will need to allocate significant resources for CAPEX investment, providing for new spending on most aspects of equipment, from the core to the edge, including customer premises gear, management systems and handsets, it will serve them to plan the new network design carefully and ensure compatibility of all their network components whether it happens to be between new device and applications and legacy equipment, or compatibility between their network and their partners’ network, well in advance.
A complete outline for the different test phases of an IMS network includes the following:
Research and Development: Test vendors are scoped and selected to assist with prototyping services. Preliminary regression and black-box/functional testing are performed. In this phase, test vendors should help service providers reduce their time to market.
Quality Assurance: Regression testing continues for quality purposes. However, in this stage, test vendors also conduct load and stress testing to determine the highest capacity of calls or sessions.
Production: This stage is perhaps the most crucial as services are actively deployed. Testing must continue to conduct load/stress testing as interoperability testing begins. This is a key stage because interoperability is one of the most important components of IMS deployment success.
Trial: Interoperability testing continues to be a crucial part of this stage. Functional testing is performed continuously throughout the lifecycle.
Deployment and Maintenance: The last two cycles are fundamental to a rollout. Active and passive monitoring and ensuring of quality of service (QoS) are essential to maintaining customer satisfaction, and test results must ensure these. As IP services become more complex, it will be essential for service providers to have a comprehensive solution for monitoring and maintenance in place.
The onus will be on telcos to insist that all participants acknowledge the critical role that protocol simulation and systems testing will play in the successful design, evaluation, roll-out and day to day operation of IMS-compliant networks, their infrastructure and equipment and the services they deliver. Telcos and service providers may find it difficult to accurately prototype and predict the effects of IMS traffic on their production networks, and may be confused when assessing interoperability of IMS-based equipment with other IMS-based equipment, either on their own networks or with partners. This will make it difficult to troubleshoot handshaking, media conversion and synchronization issues, and in turn guarantee quality of service to their customer base. The reality that IMS protocols are still evolving and will continue to do so for years, compounds the issue.
IMS implementation will be anything but a “slam-dunk,” with several complex challenges to overcome.
Perhaps the ultimate interoperability challenge that IMS poses involves mobility. A very simple single example would be a call made from a roaming mobile handset in your network. The handset will have been tested for interoperability within its own network, but not yours, and it might not behave like the mobile handsets that you have tested. Mobile providers will need to focus on SIP (Session Initiation Protocol) signaling and look at the protocol stacks within the network to ensure that in addition to being correctly and accurately implemented, they are fault-tolerant. Besides handling a session that happens to use one particular flavor of SIP properly, a network must be able to handle multiple flavors of that protocol correctly and behave gracefully if mismatches occur. If the network can’t accomplish this feat, it could drop calls and sessions or experiencing service interruptions, all resulting in frustrated users. Kind of a tall order that must be filled to be considered primed for roll-out, right?
As already discussed, the IMS protocols are evolving and in flux. SIP, a mainstay VoIP protocol, is at the front of the evolution pack and IMS SIP is not just plain old SIP. It’s SIP on steroids, so it adds complexity. The IMS specifications have imposed additional requirements on the base SIP implementation in order to accomplish the task of IMS session management correctly. These new SIP enhancements are usually at the center of interoperability issues experienced when getting IMS devices to talk to each other. There are many interpretations on how they should work, and there is much discussion concerning their implementations within the IMS development community. This makes testing and stressing of different formats of SIP headers, procedure options and variations of handshake methods critical during the first stage of deployment. Signaling exchanges have to occur properly, errors and retries need to be responded to in a timely manner using agreed-upon parameters, and attempts at invoking unsupported or unknown features must be handled cleanly and gracefully.
The last, but not least, thorn in the side of service providers when it comes to interoperability issues are network dependencies. There will be numerous entities in the IMS network performing various tasks that need to understand SIP. These entities will almost certainly not be supplied by the same vendor, and will thus have different SIP implementations. Their evolution paths will not necessarily be identical or in lock-step: the vendors will supply software upgrades and enhancements at a different rate, and possibly containing different functionality. Extra care must be taken to make sure that every time a modification is made to an entity, either in the form of an upgrade or a change in configuration, that the existing network functionality isn’t affected by the change. A series of regression tests should be executed after every change to ensure interoperability wasn’t affected by the change. Network equipment and software that have been tested with numerous vendor IMS products to validate their ability to drop into new networks with minimum interruption will go a long way toward minimizing network dependencies, achieving interoperability and provide the telco with underlying confidence that their network will support a mixed, multi-vendor environment. The key is to continually test with an updated regression suite so that the introduction of changes into the network, from diverse sources, does not produce unexpected results due to dependency conflicts.
At the end of the day, the promise IMS holds for consumers is to receive an application-rich, better, higher quality of experience, and to enjoy advanced telecommunications using any device, from anywhere with any type of media. For businesses, benefits will take shape as unified billing, negotiated bundle deals and newly-integrated services with better QoS and service level guarantees.
For Telcos to cash in, the key elements for IMS testing success include having as much real equipment as possible and driving the testing gear out toward the edge as much as possible. Service providers will own the competitive advantage if they choose test gear that emulate the edges as closely as possible and in as many different profiles as can be imagined, and have the ability to generate arbitrary traffic from any point on the network, as well as thoroughly analyze traffic at the end point. In other words, having tools in the lab that can emulate as many different flavors of SIP providing many different services will result in a faster response time, less downtime and provide faster times-to-market for new services. The tools should then be used in a continuous test cycle in a lab where all proposed changes, including software revisions, data migrations, bandwidth management and security, are tested using a well-maintained suite of test cases before implementation. Only then can you be guaranteed that the new services will operate as predicted and won’t affect existing applications.
Pierre Lynch is Director of Wireless Strategies for Ixia. For more information, visit the company online at http://www.ixia.com.
Today @ TMC
ITEXPO West 2012
October 2- 5, 2012
The Austin Convention Center
The World's Premier Managed Services and Cloud Computing Event
Click for Dates and Locations
Mobility Tech Conference & Expo
October 3- 5, 2012
The Austin Convention Center
Cloud Communications Summit
October 3- 5, 2012
The Austin Convention Center
Jive Communications by LogMeIn has Signed on as a Platinum Sponsor for 2019 Collocated Events, ITEXPO and MSP Expo