 |
|
|
Life on the border can be rough. Although borders between countries are invisible lines, wars are often fought over them, with people unfortunate enough to live in the border area often innocent victims. While the borders between cellular systems are rarely fought over, people that use their cellular phones in border areas may experience more problems with dropped calls and fewer successful incoming calls. Engineers have sought to minimize these problems, through coordinating the parameters of their system with neighbors. Standards are playing a significant role in alleviating these problems, both in improved radio interfaces and through enhancements to intersystem operations (our old friend the TIA IS-41 standard).
Border problems arise because it is impossible to completely synchronize a cellular phone with its controlling cellsite. The phone may think that it is registered with one cellsite, while the network believes that it is registered with another. This is compounded by features (bugs?) of various radio interface standards that stem from design decisions that, in retrospect, were faulty. One example is the “Stack of 4 SID” problem, present in remaining older phones that conform to IS-3 Revision C or earlier (rather than the more recent EIA/TIA-553 standard). This design feature attempted to minimize registrations by mobiles in a border area by keeping track of the last four system identifiers (SIDs) that the mobile had registered in, and only registering when a fifth SID was encountered. Unfortunately, the network was unable to track the list of four SIDs for each mobile and, even if it could have, it would have been unable to determine which of the four systems to p age in for an incoming call.
One fundamental characteristic of cellular systems – handoff – can in fact be considered a border cell problem. As a cellular phone moves around, the signal strength in the current serving cellsite weakens, requiring a handoff to another cellsite. On systems that support IS-41, the handoff can be to a cellsite connected to a neighboring system. Handoff is generally not seen as a problem, but as a challenge to be overcome. The same should be true of other border cell problems – they may occur less frequently, but most of them have solutions.
Even handoff has its own border cell problems. Until IS-41 Revision C, it was not possible to perform an inter-system handoff for an alerting (“ringing”) mobile. Consequently, a call could be lost due to low signal strength before it was answered.
The largest class of border cell problems are based on idiosyncrasies of mobile rescan algorithms. When a mobile is paged on one control channel, for example, it “rescans” for the strongest control channel before responding. Consequently, it may transmit the page response on a different cellsite from which it was paged – possibly even in a different system! This might seem crazy, but there is actually a logical reason for this. To prevent mobiles from continually rescanning while idle, they generally stick with one control channel for several minutes, or until they lose contact with the control channel. Consequently, the signal strength of the control channel at the time of a call may no longer be the best available. Performing a rescan ensures that a call starts with the best possible signal quality. This creates a problem for a cellsite on a system that receives a page that it did not initiate.
There are several solutions to this problem, either proactive or reactive. The proactive approach is to use a relatively new IS-41 message (known as InterSystemPage) to tell a neighboring system to page the mobile. Consequently, if a page response is received, the system knows how to respond to it, plus the coverage area of the page is increased. The reactive approach is to only page in one system, but to react to unexpected page responses by sending an IS-41 UnsolicitedResponse message to neighboring systems, hoping that one of them will recognize it. A hybrid approach is to tell neighboring systems to listen for a page, but not actually to initiate a page themselves. None of the solutions are perfect, as all add network load, if not paging load, but they do reduce the fraction of calls that fail due to the rescan problem.
It is not just mobile terminations that cause border cell problems due to rescans. Registrations by analog mobiles compliant to any version of IS-3 (i.e. prior to the EIA/TIA-553 standard) can cause the wrong location to be recorded for the mobile and originations can cause a similar problem. However the problem is initiated, the result is the same – attempts to call the mobile fail because the network is paging in the wrong location.
These border cell problems, and others, were first addressed in the TIA document TSB-65, which was an addendum to IS-41 Revision B. Further enhancements were incorporated in IS-41 Revision C.
There are several other border cell problems, apart from the rescan problem. One is that due to the nature of radio communications, the same message from a mobile can be received by several cellsites. While neighboring control channels should always use a different frequency, cells just one step further away often use the same frequency. The network reacts as if two messages had been received, often resulting in the mobile location being recorded incorrectly (similar to the rescan problem). This problem is minimized by a ‘digital color code’ that adds another level of discrimination beyond frequency, and can further be combatted by comparing the signal strength of the two messages. TDMA digital cellular (IS-54, IS-136) has further reduced the likelihood of this problem occurring by increasing the number of color codes available, and the problem cannot occur with CDMA (IS-95) systems.
A relatively new border cell problem occurs with authentication. This anti-fraud technology relies on the ability of a cellular system to challenge a mobile with a random number. Because of border cell problems, one-quarter of the random number is echoed back to the cellsite by the mobile. However, this does not guarantee that the mobile did not pick up a random number from a neighboring system that is three-quarters different (which would result in a false authentication failure). If a mismatch in the echoed portion occurs, the receiving system needs to determine the full correct random number, which may have been transmitted by a neighbor. Fortuitously, IS-41 Revision C contains a message that allows the coordination and exchange of these random umbers.
The efforts of radio, network and standards engineers are making life on the border steadily more uneventful. People that use their phones in these areas (such as around densely populated parts of the Great Lakes) should notice a continual increase in the quality of service that they receive. Also, newer phones, and especially digital phones, should experience fewer of these problems.
© – Copyright