DGNA Capacity Planning
GeoFence DGNAs depend on the position reports (LIP) of the terminals: a fence change is only detected when a new position arrives, and every fence change usually causes two DGNAs (one for leaving the old fence, one for entering the new fence). Both the LIP reports and the DGNA acknowledgements use the uplink of the TETRA node the terminal is registered on. The example below shows how to check whether a planned configuration is feasible.
Version
The DGNA throughput described here applies to v3.0.9 / v3.1.2 and later. Older versions process DGNAs strictly one after another (approx. 8–10 DGNAs per minute) and can fall behind by several minutes under load.
Network inputs (example)
| Parameter | Value | Unit | Notes |
|---|---|---|---|
| Number of nodes (base stations) | ? | nodes | Single-carrier nodes: TS1 = MCCH, 3 traffic channels |
| Average terminal speed | 25.0 | km/h | Typical vehicle speed in the planning area |
| LIP reporting interval (time trigger) | 30.0 | s | Configured periodic position report |
| LIP distance trigger | 250 | m | Report after this distance travelled; enter 0 if not used |
| Total terminals | 3,000 | terminals | All terminals registered in the network |
| Active position-reporting terminals | 300 | terminals | Terminals sending LIP reports |
| Average group call share | 30 % | share of time | Share of time terminals spend in group calls (> 30 % = talk-heavy) |
Planning assumptions (to be verified on the target system)
| Parameter | Value | Unit | Notes |
|---|---|---|---|
| Target uplink utilisation | 80 % | share | Headroom for status messages, other SDS, registrations, call set-up and DGNA acknowledgements |
| Load factor of busiest node | 4.00 | × | 1.00 = reporters evenly distributed; 1.50 = busiest node carries 50 % above average; 4.00 = many terminals are served by the same node |
| Group timeout | 5.0 | s | Average wait until call ends (terminal in a call) |
| DGNA delay for terminals in a call | 15.0 | s | Assumes DGNA takes effect after the ongoing group call has ended |
| Uplink slots per LIP report | 2 | slots | Random access plus reserved data slot, including occasional retries |
| DGNA duration (terminal idle) | 1 | s | Median measured on TetraFlex |
Position reporting load per node
- At 25 km/h, the 250 m distance trigger is reached after 36 s. The 30 s time trigger is therefore decisive, and a terminal moves approx. 210 m between two reports.
- 300 reporting terminals generate 10 LIP reports per second in the whole network.
- The MCCH uplink of a single-carrier node offers approx. 17.6 slots per second. At 80 % target utilisation and 2 slots per report, one node can handle approx. 7 LIP reports per second, i.e. approx. 210 terminals reporting every 30 s.
With a load factor of 4.00, the busiest node carries four times the average number of reporters (at most all 300):
| Nodes | Reporters on busiest node | LIP reports/s | Uplink utilisation | Result |
|---|---|---|---|---|
| ≤ 4 | 300 | 10.0 | approx. 113 % | ✗ Not feasible |
| 5 | 240 | 8.0 | approx. 91 % | ✗ Not feasible (no headroom) |
| 6 | 200 | 6.7 | approx. 76 % | ✓ Feasible (tight) |
| 8 | 150 | 5.0 | approx. 57 % | ✓ Feasible |
| 10 | 120 | 4.0 | approx. 45 % | ✓ Feasible |
Rule of thumb: one node should not serve more than approx. 210 terminals at a 30 s interval (approx. 105 at 15 s, approx. 420 at 60 s). If more reporting terminals gather at one node (e.g. depot, staging area), increase the LIP interval or add a carrier.
DGNA throughput
- With 30 % group call share, 70 % of the DGNAs take effect after approx. 1 s, 30 % only after the group call has ended (approx. 15 s). This results in an average DGNA duration of approx. 5 s.
- GroundControl sends several DGNAs at the same time (setting
Dgna:MaxParallelof the Internal Event-Processor, default 4, max. 16 = TetraFlex limit of simultaneous DGNA instances):
Dgna:MaxParallel |
DGNAs per minute | Fence changes per minute (2 DGNAs each) | Minimum time between two fence changes per terminal (300 reporters) |
|---|---|---|---|
| 4 (default) | approx. 46 | approx. 23 | approx. 13 min (approx. 5.4 km at 25 km/h) |
| 8 | approx. 92 | approx. 46 | approx. 6.5 min (approx. 2.7 km) |
| 16 | approx. 185 | approx. 92 | approx. 3.3 min (approx. 1.4 km) |
Higher values of Dgna:MaxParallel only help if the terminals are spread over many nodes, because the DGNAs of one node share its MCCH (part of the 20 % headroom).
Feasible and not feasible loads
Feasible
- Up to approx. 210 reporting terminals per node at a 30 s interval (at least 6 nodes in the example above).
- Terminals change fences on average no more often than every 13 min (default) or every 3.3 min (
Dgna:MaxParallel= 16), e.g. large fences along a route. - Short bursts, e.g. 50 vehicles entering a fence together: 100 DGNAs are completed after approx. 2 min (default) or approx. 35 s (16 parallel). Outdated DGNAs are skipped automatically, so a backlog does not grow without limit.
Not feasible
- All 300 reporters on 1–5 nodes with a load factor of 4.00: the uplink of the busiest node is overloaded, position reports and DGNA acknowledgements get lost or delayed.
- Fences smaller than the distance travelled between two reports (approx. 210 m at 25 km/h and 30 s): the terminal may pass the fence without a report inside it, so no DGNA is triggered.
- Many small, adjacent fences that all terminals cross continuously: in the worst case (fence change at every report) 300 terminals cause 20 DGNAs per second, which would require more than 100 simultaneous DGNAs.
- DGNAs that must take effect immediately while the terminal is in a group call: the group is changed only after the call has ended.
Expected delay
The time from entering a fence until the DGNA takes effect consists of the waiting time for the next position report (on average 15 s, at most 30 s) plus the DGNA duration (approx. 1 s, or approx. 15 s during a group call). Plan with approx. 15–45 s in the example above.