The buildout of 2.5G and 3G networks is under way, and consumers are showing a visible interest in using their mobile devices for more than just talking. But for subscriber numbers to explode, these services must be reliable and consistent. Thus far mobile operators have been so focused on bringing the new services to market that they have invested little in tools that analyze network performance and assure service delivery.
"Service assurance tools are rarely rolled out with the service," says Dave Gellerman, vice president of technology and corporate development at Spirent. "Instead, operators wait until the service is successful. Newer applications, such as location-based services and m-commerce, won't have any service assurance component until they achieve a scale that makes the investment economical."
This policy is already in play among some mobile services. For known high-revenue customers, such as roaming subscribers, or mission-critical applications, such as emergency facilities that use SMS messages for dispatch, mobile operators are tracking transactions and monitoring network performance closely.
"For high-revenue customers, mobile providers are adding surveillance tools now," says Boyd Williamson, wireless business manager at Agilent. "These tools are not universal, but providers are recognizing that they must monitor high-value services."
Tying Customers to Services
An overarching theme at the service management companies is the operator's need to combine network performance data with mobile customer information. The service providers don't need more alarms in the network operations center (NOC) that isolate faults on the network, they claim, but the carriers do need tools that wrap network data with the customer information. Currently "most carriers only get a snapshot of network performance, not a complete picture of how poor performance affects customer service," says Chuck Stormon, chief technology officer and co-founder at Coherent Networks.
Understanding the quality metrics surrounding services delivery will be key to providers charging additional dollars for new services, but impossible with traditional network management tools, says Deepak Swamy, senior vice president of marketing at Trendium. "In the wireless environment, to discover if the service was delivered and how well it was delivered, carriers must pull together data from multiple sources," he says. "They need to know if an entire video file was downloaded or if an m-commerce transaction was successful. These measurements are much different from measuring call completion or call blocking and require a broader set of measurements from a wider range of elements."
Today's network monitoring is done by sampling the network, or taking measurements at specific times of the day. The providers rely on network performance metrics from single-system element management systems (EMSs) or network management tools, such as HP OpenView or Micromuse's NetCool. These traditional NOC management tools don't, however, provide information about which customers lost service or what traffic generates the highest revenue.
"Providers don't know if their traffic includes 1,000 short messages or 10 high-value transactions," says John Konczal, vice president of product marketing at Telution. "They need to know which customers they have commitments to and which transactions have a higher priority."
Such real-time knowledge of network, customer and service performance is rare within mobile operators' NOCs. But the providers can furnish monthly performance reports for customers. To create these reports, many providers use a labor-intensive, manual method that they offer only to select customers. Every month a member of the IT staff must go into different EMSs and pull out statistics that relate to specific customers. These reports tie network elements to the customer and provide metrics that are imported into an Excel spreadsheet and used to create a summary report. For one or two customers, this manual method is manageable; for hundreds, it's impossible. Many of the service assurance tools automate this process, making the practice more scalable, and providing a broader view of the customers and network (For more on monitoring tools, see "Monitoring the Wireless Network").
The more dynamic wireless networks, though, present more challenges than fixed environments. The most obvious difference is the subscribers' mobility, which makes qualifying customer service based on network performance difficult for operators.
"Mobile providers can't identify a port with specific services. It's essentially impossible to track failures to what subscribers are affected," says James Storey, chief technology officer at TCSI. "Instead, providers should track the location of their high-revenue customers' handsets and improve network performance based on these valuable customers' needs."
Drilling to an even deeper level, or tying network performance to a unique customer rather than a set, is also problematic for operators, explains Matt Standish, product marketing manager at Concord. "Mobile providers must gather information from up to seven parts of the network and compare it to a customer database. On a wireline network, MPLS would identify the customer by IP address. A PDA or mobile phone doesn't have a specific IP address, so it's difficult to prioritize a customer's traffic across the network."
Another core challenge of monitoring performance in a wireless environment is that most next-generation services rely on multiple shared resources. Service delivery depends on the existing radio equipment, voice switches and service nodes, as well as additional elements such as content caches; email, WAP or application servers; SIM cards; provisioning systems; and many other components now being added to the wireless network.
"The network infrastructure has become complicated with many more elements playing a part in delivering services," says Trendium's Swamy. "Application components, server and system components, and network and business components are now a vital part of the service delivery chain."
Reality Check for
Wireless SLAs
The new components in the service delivery chain, as well as the additional partnerships required to offer content services, have prompted debate on the merits of tight integration in the wireless OSS. Does a tightly integrated OSS allow carriers to improve provisioning, ensure delivery and allow SLAs? Or does it limit business process changes and future service options?
Many carriers have been demanding more open, modular systems, while the OSS providers have countered with the benefits of tightly coupled systems. Coherent's Stormon predicts that wireless OSS integration will be adapted piecemeal to accommodate new services and technology, but will become more tightly integrated as applications grow more sophisticated.
"At first mobile operators will put in network monitoring systems, but they won't be attached to the OSS," he says. "They will add a new billing system, but not attach it to the network. Before they are finished rolling out the GPRS or 3G network, they will realize the critical need for integration between the BSS, OSS and network monitoring."
The next-generation services will demand that operators create methods to share network information with the OSS, says Telution's Konczal. "Content delivery will require provisioning on every single transaction, and each time the parameters could be different depending on the wireless device, the user's location and the type of content."
Tightly joining the OSSs and network could lead to new SLAs for business customers and content partners. As wireless service becomes more mission-critical within businesses and content services begin generating revenue, these customers and partners will demand SLAs that guarantee quality of service and exceptional performance.
"To date wireless carriers have made no commitments, because they couldn't stand by them," says Konczal. "New business models will have to support commitment-based offers for corporate customers."
Large enterprises that have major contracts with mobile providers will expect certain guarantees. These customers will be evaluating percentage of calls blocked, calls dropped, quality of service, provisioning times, and help desk response times. They may also demand that their executive management team receive a higher level of performance than other employees or that field personnel be guaranteed reliable service when interacting with the corporate network.
"VIP customers will need more
granular performance details and cross-monitoring," says Agilent's Williamson.
The more interesting-and complicated-SLAs will be between mobile providers and their content and retail partners. M-commerce and content delivery services will require contracts between all the players in the service supply chain. These SLAs may surround completion time, security, connectivity and transaction success rates.
Initially, the SLAs will be relatively simple. "Agreements will address service availability, appropriate performance, throughput and transaction time," says Spirent's Gellerman. "Once the services become more sophisticated, we will get more complicated SLAs."
Provisioning Commitments
Many of these upcoming contracts will not relate to performance but provisioning. Content partners may require a commitment to enable the service within a particular time, to customize the mobile phone interface around the content provider, or to place the content in certain areas on the phone. These contracts address the look and feel of the content, how it will be assessed and whether it requires revenue sharing.
"Most wireless infrastructures are built for the consumer," says Konczal. "When customers sign up, they are provisioned directly to the network element. Partner contracts are completely different. Multiple points in the business touch this order, creating more of an assembly-line architecture. Wireless carriers will have to turn content contracts into information that can be provisioned within the wireless infrastructure. Our customers are struggling with an OSS environment that is not prepared to handle this type of partner information."
And finally, operators should use their OSSs to help them decrease operating and capital expenses and increase revenue. "The OSS can help mobile providers cost-effectively build their networks by planning for capacity. Their goal should be to create new services, or new profit margins, without creating additional expense," says Coherent's Stormon.
Having broad experience in the wireline world, performance-monitoring companies are applying their network management knowledge to the wireless space. These companies are monitoring the backhaul links between the base stations, BSCs (base station controllers) and switching centers; however, none are attempting to monitor the RF link, which has the highest failure rates and is the most difficult to monitor. The companies are approaching wireless providers with differing strengths, and each has a unique spin on how it can improve operators’ visibility into the network. As one of the larger providers, Agilent is tackling service assurance from multiple directions with different offerings. “No single tool can solve the problem,” says Boyd Williamson, wireless business manager. From a straightforward network management approach, Agilent looks at congestive noise and traffic patterns. Products also perform active tests throughout the network and the service activity. To test email, Agilent would launch a request through the network by sending an email, retrieving the email and looking at service availability and traffic latency. These types of ongoing tests monitor how well the individual services are operating. Another aspect of the tests is following the transactions across the network, using sampling, to discover users’ individual experiences.
“We break the network up so we can monitor at the network, connectivity, service and individual level,” says Williamson. Ai Metrix has an integrated suite of tools for system integration, service activation, service assurance and inventory management. Within service assurance, its software monitors network events, messages and order fallouts. Monitoring these events enables providers to understand how network resources and connectivity are related to customers and services.
“Today mobile providers are focused on pure fault management. They look at hard alarms and broken cards,” says Mark Wyman, vice president of product management. Ai Metrix monitors the trunks of multiple technologies in real time to understand what is not working in specific areas. It then combines this information with facilities information to provide a more integrated view of the network. Coherent monitors the network layers. At the bottom layer, from the switches and cell towers, operators are dealing with fraud, network capacity and QoS issues with dropped calls. At the second level, carriers are running heterogeneous networks, having started with TDMA or CDMA and layering GPRS. Here, traffic management and performance become difficult. The third level is where wireless carriers become part of the global network and have to deal with interoperability, interconnection and termination traffic.
“Each layer has a different set of concerns and uses different signaling interfaces,” says CTO Chuck Stormon. “We can provide analysis, visualization and traffic alarms, because we understand the variant protocols of the SS7 network and can correlate the information and create a view of end-to-end behavior.”
Concord is leveraging its fault isolation and network performance experience, placing particular emphasis on monitoring all the technologies that support wireless service. The company doesn’t look at the radio access side, but manages all the equipment behind the base transceiver stations to understand IP, ATM and optical activity.
“Concord helps identify the relationships across the network,” says Matt Standish, product marketing manager. “Operators shouldn’t add bandwidth if the bottleneck is a router, a router card or a downed node.”
The company also supports capacity planning, which will be important as content-rich services, such as videos, photos and other downloads, become more prevalent. “We can help create bandwidth on demand for applications that will draw bandwidth on the trunk-side infrastructure,” Standish says. “These services will need to be managed closely to understand capacity, performance, quality of service, and availability in a wireless environment.”
TCSI launched its service assurance for wireless carriers in February 2002. The company aims to help providers understand the business impact of events and faults that disrupt wireless service. Its software collects and integrates fault, performance and topology information from the network and OSSs.
“Providers can focus on high-revenue customers by prioritizing alarms for resources in the NOC,” says CTO James Storey. “The software looks at performance analysis and trends and can gauge performance at cells, understanding the trends of poor quality.” By gathering information from the network elements, the network management layer and the service management layer, TTI Telecom analyzes service availability and performance. The company’s service management offering has multiple modules including fault performance, configuration and provisioning, performance management and CDR analysis.
“We try to understand what is occurring on the network from a service and customer point of view,” says Danny Itzigsohn, marketing manager for wireless solutions. “Several faults can be critical from a network point of view but not disrupt service. We want to know what exactly went wrong with the customer’s service and provide that useful information to the operator.”
Spirent targets fault isolation and performance management along the backhaul network, or from the base station into the network. Its tools provide network surveillance and isolate faults occurring in access to the PSTN or backhaul.
“In a GPRS environment we gather performance data from the Gateway GPRS Service Node and the Serving GPRS Support Node,” says Dave Gellerman, vice president of technology and corporate development. “Gathering performance metrics from these data devices is similar to other data devices, such as ATM switches and IP routers, but quite different from gathering data from the transport network or mobile switches.” Ensuring service delivery and understanding network and OSS capabilities is a priority for Trendium. Its software uses probes to gather information from network elements, EMSs, radio elements and gateways; application monitors gather information from servers and systems; and agents collect data from OSSs involved in the service delivery chain.
“Providers need a model of how service is delivered from the network infrastructure and OSSs,” says Deepak Swamy, senior vice president of marketing. “They could deduce that information from fault information or events on the network, but that method is not as informative as an overarching model and analytics.”
Trendium has been working with a U.K.-based 3G operator to provide cause-and-effect analysis and service prediction. “The operator has multiple layers of business processes that make up the service,” Swamy says. “We gather metrics from the GSM and radio networks, the infrastructure components, and the services from the CRM and provisioning systems.
“We use these metrics to model the service and show the dependencies on each of the underlying business processes or server components. The model shows the logical topology of how every element and system regarding the service relates. The top level shows the 3G applications and server farms. Below that are the underlying GPRS infrastructure and business processes relationships to the provisioning system and the individual network components.”
Email this article
Add a comment
Printer version
Order reprints
RSS Feed
Bookmark article