White Paper - Emergency Call Handling in VoIP Networks
Emergency Call Traffic Management (continued)
When an emergency event occurs, people not only make emergency calls but also try to ring their friends and relations to assure them they are OK or to find out if they are affected. Of course, if the call does not get through, they repeatedly try again, amplifying the network congestion problem. Depending on the scale of the event, this traffic can be confined to localities or spread nationwide.
In the midst of this huge spike in traffic, key personnel will be trying to organise an effective response. If their calls are not effectively prioritised then they will not get through. The new challenge is to allow this traffic to get through at the expense of none essential, personal calls.
The main SIP RFC 3261 does not support the required resource prioritization; however, there is a recent supplementary RFC (RFC4412) designed to extend SIP's capability in respect to call prioritisation and the allocation of resources. This extension is intended to cover a number of service scenarios including the requirements of emergency call handling. The RFC defines new SIP header fields that allow a device to request that the call is treated by downstream elements as a high priority call. If this scheme is adopted it will allow elements such as SBC which are managing session admission to pass these calls in preference to others.
The problem is particularly acute at the network interconnect points, say between a mobile and wireline service provider. The interconnect point will probably come under severe strain from personal callers. If the right priority marking is attached to the SIP messages, then an interconnect session border controller can readily deal with the traffic, rejecting lower priority calls.
There are a number of schemes that can be used, but the key need is to allow the automatic prioritisation of traffic, and the appropriate selective rejection of new calls.
Newport Networks' 1460 SBC employs and advanced scheme where emergency calls are allocated an initial reserved capacity and the traffic management works to keep that reserved capacity clear for new emergency calls using a sliding window philosophy. For example, in the figure above, five low priority calls are permitted and space is reserved for two new emergency calls. Once the limit of five low priority calls is reached, no further low priority calls are admitted. However, emergency calls are admitted into the reserved bandwidth. When a normal call ends, the system reclaims the reserved capacity for two new emergency calls. The system achieves this by reducing the normal call count when calls end. Thus, if the total capacity is seven and one emergency call is active, the system will limit normal calls to a maximum of four thereby preserving the new emergency call reserved bandwidth of two.
This concept can be extended to support a key personnel band, which is treated in a similar manner to emergency calls but at a lower priority. The aim of this method is to always ensure that resources exist for emergency calls and capacity is automatically allocated as the event unfolds, sometimes very rapidly and beyond the ability to control through a manual response by network operators.
Summary
Effective handling of emergency calls is vital for all VoIP networks. Identifying emergency calls, prioritizing them and delivering them to the emergency operator is a key requirement. The protocols and procedures needed to enable this service exist and can be delivered today. Newport Networks 1460 session border controller supports prioritised delivery of emergency calls today and Newport is committed to supporting the emerging standards that enhance the effective delivery of emergency calls. ■
|
IPsec and VoIP
