Handoff prioritization and trunking terms

Two schemes for prioritising handoff

Dropping an ongoing call is worse than blocking a new call, so handoffs are given priority:

1. Guard channel concept

   • A fraction of the channels in each cell is reserved exclusively for handoff requests; new calls can use only the remaining channels.
   • Reduces dropped calls, but lowers total carried traffic because guard channels may be idle. Dynamic channel assignment makes it more efficient.

2. Queuing of handoff requests

   • There is a finite time interval between the signal reaching the handoff threshold and falling to the minimum usable level. During this time the handoff request is placed in a queue and served as soon as a channel becomes free.
   • Reduces forced termination, but cannot guarantee zero drops because the waiting time is limited.

 Signal level
   │ \
   │  \ ← handoff threshold: request queued
   │   \
   │    \ ← minimum usable level: call drops
   └──────────── time
       |<-->| queuing time available

Trunking terms

• Grade of service (GOS): a measure of the ability of a user to access the trunked system during the busiest hour; usually given as the probability that a call is blocked (Erlang B) or delayed beyond a given time (Erlang C). Example: GOS = 2 %.
• Traffic intensity: the measure of channel time utilisation, i.e. the average channel occupancy, measured in Erlangs. For one user, Au = λH; total offered traffic A = U·λ·H. One Erlang = one channel occupied continuously.
• Average holding time (H): the average duration of a typical call. Example: H = 3 minutes.
• Blocked call (lost call): a call that cannot be completed at the time of request because all channels are busy (congestion). In Erlang B systems blocked calls are cleared; in Erlang C they are queued.