SS7 Monitoring: How to Make the Most of Interconnects

The interdependence of carrier networks is only becoming more complex. In the past, most calls originated and terminated within a single network, and relations between companies were based largely on trust and rough estimates of traffic ratios. Today, voice calls and other messages often traverse multiple independent networks, and telephone operations function properly only to the extent that the networks interconnect well.

These days, interconnect management is critical to a service provider’s financial success. To ensure the satisfaction and loyalty of its own customer base, a carrier must do everything it can to guarantee the performance and good faith of the other carriers that handle its calls. Likewise, a carrier often handles so much traffic from outside its network, it must do so in a way that recovers costs and generates revenues.

This is no easy task in a complex technical environment, with so many participants in the service chain and so many different kinds of traffic. The challenges are diverse, and carriers have yet to settle on their best strategies. One method is to use information inherent in Signaling System 7 (SS7) to collect and synthesize data in real time and analyze it to verify that service level agreements (SLAs) are being met.

Case in Point

Jack, who lives Houston, wants to call and propose to Jill, who studies law in Boston. His call might pass from Master Mobile, his wireless service provider, through a CLEC, a long-distance carrier, an ILEC in Boston, and then on to Jill’s own wireless network. But in this case, the ILEC’s network is congested, and Jack gets a recording. Master Mobile loses revenue because Jack only pays for calls that succeed. And who is Jack likely to blame? Master Mobile, even though his carrier has done everything right.

Conversely, Jill considers Jack a nuisance and pays for the caller ID services of FaultFree Wireless to avoid his calls. But Jack’s calls traverse the network of a long-distance carrier that does not use its SS7 signaling to deliver calling-party numbers. Consequently, Jill’s caller ID service is unable to display Jack’s number, and Jill cannot screen out his calls. Who is she going to blame? The faultless FaultFree. She cancels caller ID, or perhaps even the wireless subscription itself.

There is a growing trend among service providers toward including SLAs as part of their interconnect contracts to handle such problems. For example, an SLA might specify the precise percentage of calls that can come into a carrier’s network without the calling party number. When the actual percentage is higher, the agreement would allow the receiving carrier to assess penalties on the carrier that transmits the calls. Similarly, an SLA might impose acceptable “answer/seizure” ratios, requiring that a certain percentage of calls passing from one network into another actually make it through that network.

However, an SLA is only effective if the carrier can measure performance (See Figure 1). How do they monitor interconnect traffic to determine answer/seizure ratios or the percentage of calls that include the calling party number? This information can be extracted from the basic message signaling units (MSUs) that constantly pass back and forth between the transit switches of SS7 networks. Although many transit switches are capable of generating the relevant call detail
records (CDRs), not all can. Even where it is possible, many carriers may turn off the capability because it impacts the performance of the switch. Alternatively, carriers can deploy systems that build synthetic CDRs from MSUs at the point of interconnect. These systems can be customized to include advanced filtering techniques, therefore synthetic CDRs are often richer and more pertinent to the relevant service issue than switch-generated CDRs.

As the trend toward measurable SLAs reaches critical mass, they will become competitively indispensable, as will the systems that monitor them. For example, if carrier A signs an SLA with carrier B, A’s ability to maintain the agreement may well depend on the performance of carriers X, Y and Z. Hence, A will have a strong incentive to enter similar agreements with these companies.

Monitoring systems can also have positive service quality implications in networks with switches that automatically generate CDRs. Switch-generated CDRs are not delivered in real time, but rather at intervals of 15 minutes or more. In several contexts, the timeliness of CDRs is very important. For example, real-time CDRs can make a significant difference when carriers evaluate their options for routing calls at international gateways.

Suppose a call originates in Britain and terminates at a LEC in the United States. The British operator can route the call via a U.S. long-distance carrier at, say, 5 cents a minute. However, it may be cheaper at this particular hour to use an international carrier in Japan. Since it is the middle of the night in Japan, the Japanese carrier has a lot of spare capacity, and offers its service at 3 cents a minute. Oftentimes local carriers can choose between the discounted services of several international companies in what amounts to a spot market for spare capacity.

Routing Decisions Based on QoS

Suppose the British operator chooses the discounted network services of international carrier X because they are cheaper than those offered by Y and Z. Something then happens on X’s network that prevents it from completing a significant percentage of calls. To maintain high quality of customer service, the British company may well want to reroute its traffic via Y. In order to minimize the number of lost calls, the British operator needs in real time the relevant CDRs that include information on answer/seizure and answer/bid ratios. If transit switches only generate CDRs every 30 minutes, it may take the company as long as an hour to discover the problem, which is a longtime to be sending calls through a malfunctioning network. The company can solve this problem with a commercial application that generates real-time CDRs. Such applications offer an additional benefit by allowing operators to reduce the processing burden on their switches. The processing power required to generate CDRs can now be assigned to the switches’ primary function of switching.

Accuracy in Compensation

When local competition arrived, LECs really weren’t set up to deal with the complexities of reciprocal compensation at the local level. Since their transit switches did not generate CDRs, companies lacked the detailed records required to accurately bill for trunk access and to verify the trunk access bills they were receiving from other LECs. Early on, some carriers adopted an “I won’t charge if you won’t” approach, but as the discrepancies in traffic were often large, the system could not last. This method was replaced by rating factors, where billing was based on rough estimates of traffic ratios. In making these estimates, carriers found it particularly difficult to accurately identify different kinds of traffic in order to apply different rates.

In recent years, two important developments have introduced precision and fairness into the billing process: switch vendors have improved the capabilities of their switches, and test and measurement companies have marketed link-monitoring systems that generate the relevant CDRs in both directions. These records include where a call originates and terminates, on which trunk, the start time, end time and other relevant call information. The CDRs can then be fed to billing applications or used to verify interconnect charges.

Link-monitoring systems are especially useful for identifying the precise types of traffic that traverse interconnects, a crucial capability that ensures that carriers get paid fairly for calls that command different rates. For example, companies that have deployed such systems have discovered that standard estimates of local-toll ratios—PLU (percent local usage)—are often wrong, with errors usually in favor of the carriers paying the bill. Such errors can be very costly.

In another example, LECs with monitoring systems have discovered significant errors in the standard estimates that underlie the PIU (percent interstate usage), a rating factor that can dramatically affect the access charges LECs receive from long-distance carriers. These errors usually benefit the long-distance carriers, because the factors overestimate the amount of inter-state traffic compared to intrastate traffic. LECs are often entitled to charge more for handling intrastate calls because these calls attract state tax, whereas interstate traffic does not. SS7 monitoring systems can generate CDRs that include calling-party numbers, which enable LECs to determine the states in which calls originate.

Monitoring systems can also help carriers solve a related but more complex problem involving PIU and PLU arbitrage. Carriers sometimes pass traffic through MF (multifrequency) trunks before delivering it to the next carrier. Since MF signaling does not include the calling party number, monitoring systems can’t generate the CDRs necessary to determine whether a call is interstate or intrastate. But LECs can program these systems to determine what percentage of incoming calls have traversed MF trunks (See Figure 2). They can then use this information as the basis for strict, measurable SLAs that state an acceptable amount of incoming MF traffic, and assess penalties when the MF traffic exceeds that amount. Therefore, carriers will be less likely to use MF trunks, and LECs will be more likely to receive proper compensation for their services. Monitoring systems can help with other issues of arbitrage. Suppose, for example, a wireless provider has an interconnect agreement with an ILEC that charges a high per-minute rate for incoming wireless traffic. The wireless service provider might decide to save money by routing its calls through a CLEC that charges half that price; the CLEC then passes the traffic through to the ILEC as if it had originated from within the CLEC, and pays only reciprocal compensation rates on this traffic. The wireless provider gains by paying a lower rate to terminate its traffic, and the CLEC makes a profit on the traffic that transits its network. Only the ILEC loses out. The ILEC can now detect this strategy with system-generated CDRs, and demand payment at the appropriate higher rate. In general, link-monitoring systems can go a long ways toward resolving arbitrage issues.

Call Analysis Through Pattern Matching

Monitoring applications can build CDRs for call-pattern matching, which can be used to determine the identity of parties who either originate or receive calls in other networks. Sometimes carriers need this information in order to properly rate interconnect traffic, generate revenues and cut costs. Consider two topical examples.

Suppose an ILEC has many subscribers who frequently make long-duration calls to a single number within a CLEC’s network, and yet the ILEC receives no calls from that same number. There is a good chance the number belongs to an ISP. The ILEC makes very little from such traffic because calls to ISPs are usually rated as local. Hence, the ILEC pays reciprocal compensation fees and is never compensated in return. The CLEC, on the other hand, makes a bundle for terminating so many calls, which is why some companies have done their best to sign up ISPs.

Recently, some carriers argued to the FCC that standard reciprocal compensation rules ought not apply in this situation because ISP calls are really long-distance. Why? Because when subscribers call their ISPs, they are potentially connected with any Internet site located anywhere in the world. The FCC ruled that this point can be negotiated as part of interconnect agreements. When carriers agree that ISP calls are long distance, reciprocal compensation fees do not apply.

This is fine in theory, but how do carriers determine which calls go to ISPs? They can deploy pattern-matching applications that monitor interconnect traffic for the features characteristic of ISP calls. For example, the peak time for such calls is 9 p.m., as opposed to 3 p.m. for voice calls; and on average, ISP calls last significantly longer than voice calls. When the monitoring application determines that calls to a specific number have such characteristics, a carrier can check the number against a list of known ISPs. If the number is not on the list, the carrier can simply call that number to settle the question.

Pattern-matching applications can also help carriers monitor interconnects to detect a costly and increasingly popular method used by subscribers to avoid expensive international call charges. This involves the use of “call-back” services, which while perfectly legal in many countries, are illegal in others. This is because the incumbent carriers and/or government regard call-backs as defrauding the local incumbent of their legitimate international calls. It relies on the fact that calls from, say, the United States to other countries are usually cheaper than calls going the other way, even if they traverse the same network.

Call-back fraud has several different forms, some of them quite sophisticated. The following example is typical. Boris in Russia needs to communicate frequently with several people around the world, all of whom have different telephone numbers. But Boris always dials the same foreign number and then hangs up without waiting for an answer. He doesn’t wait because he is actually calling a “trigger” number assigned exclusively to him. A computerized agent then calls him back and offers a voice menu that allows him to enter the actual number of the person he wants to call. Boris ends up paying a lot less, and the Russian telephone carrier receives no fees for originating a completed long-distance call.

Although call-back schemes are not always illegal, the laws of most countries do not prohibit carriers from blocking calls to a specific number. But how do carriers determine which numbers to block? One test and measurement vendor has recently developed a pattern-matching application that filters the CDRs from outgoing international traffic and identifies batches of calls to the same numbers that exhibit sequence patterns characteristic of call-back schemes. A carrier can then investigate those numbers to determine whether they are indeed used by call-back vendors.

New Measurements for New Revenue

Link-monitoring systems enable carriers to manage interconnects that maximizes return for existing services. Monitoring systems can also help generate revenue from new, nontraditional sources. For instance, carriers using such systems have discovered that a significant percentage of interconnect traffic consists of SS7 messages that have nothing to do with voice calls.

Suppose Jack turns on his cell phone in Boston. Since he has roamed into FaultFree’s wireless network, FaultFree will send an SS7 query to Jack’s wireless provider in Houston, which will respond with a message confirming that Jack has roaming privileges. This happens every time he activates his phone, regardless of whether he makes or receives calls. Since these messages traverse the networks of LECs and long-distance carriers, the intervening carriers are beginning to insist on compensation. Monitoring systems provide the CDRs required to accurately bill for such traffic. Monitoring systems will also enable carriers to track and bill for SMS (short message service). SMS allows subscribers to send text messages via SS7 networks, a service that has already become popular in Europe and is bound to catch on in the States.

Large carriers are also beginning to generate revenue by supplying advanced SS7 capabilities to small carriers that lack the infrastructure to provide their customers with special services. A start-up in Wyoming may want to offer 800 services, call portability and call-name delivery, but it lacks the capital to purchase and maintain the extremely expensive databases that SS7 messages must query to enable these services. Such a carrier can purchase access to the databases of larger operators that have the necessary infrastructure in place. Link-monitoring applications enable these carriers to charge for access on a per-message basis.

IP and Future Challenges

The increase in phone usage and the advent of services such as SMS are beginning to exhaust network resources, and many carriers are adopting an SS7-over-IP transport structure to enhance capacity and speed. As this trend continues, more SS7-over-IP networks will need to communicate, and carriers are also likely to convert their interconnects to SS7-over-IP. SS7-over-IP will give rise to entirely new traffic patterns, new possibilities for network error, and new opportunities for arbitrage and fraud.

Currently, when two SS7-over-IP networks exchange traffic, the SS7-over-IP signaling must be transformed at the interconnect to circuit-switched SS7 and then back into SS7-over-IP. This wastes processing power, increases the possibility for transmission errors, slows traffic and diminishes service quality.

As the move toward SS7-over-IP accelerates, carriers will face new complexities in interconnect management, and link-monitoring systems will have to keep pace. IP networks are intrinsically less secure than circuit-switched SS7 networks. It is far easier to introduce malicious traffic into IP networks than into the closed SS7 networks of today. As a result, carriers will need systems to detect this traffic at the point of interconnect.

The transition to SS7-over-IP is perhaps the most prominent technical issue facing service providers over the next few years, many more challenges will arise from new and unforeseen technologies. One thing is certain: As technology transforms the world of consumer choice, carriers will need ever more sophisticated link-monitoring systems to maximize revenue and ensure quality across their interconnects.

John Clark is a Product Manager with Agilent Technologies, with responsibility for the successful Business Intelligence Applications, part of the world-leading acceSS7 network monitoring system solution. Clark joined Agilent in May 1997 from GPT Public Networks Ltd, where he held a number of marketing positions, including Product Manager on the System X switch evolution, and Market Development Manager in China. Clark has over 10 years of telecoms industry experience, and has had a number of articles published in leading journals. He can be reached at john_clark@agilent.com.