WiFi has a topic that is not much discussed but which has a great impact on the actual effect: the relationship between flow orthopaedic strategy and billing accuracy. Many people think that precision depends only on the calculation of the costing engine itself, and that it is not related to network management. But actually, traffic orthopaedic configurations have a direct influence on the flow data actually generated by users, and hence on the measurement results. If this relationship is not clarified, they will appear when the system is online."The billing data appear to be correct, but the actual experience and expectations of users are not consistent"The ambiguity is particularly difficult to identify.
The position of the flow plastic in the billing chain.
The user data is sent from the terminal, via AP, AC, flow orthopaedic modules, and billing modules, and finally reaches export. If it occurs before the billing module (usually in this order), the actual traffic generated by the user has been modified by the orthopaedic module.——The drop of bags, delays, speed limits all alter the original flow characteristics. The billing module counts a post-reform traffic, not a flow actually generated by user terminals.
Direct impact of speed-limiting strategies on traffic statistics
Speed limits are the most common means of flow orthopaedic.——For example, the 5-mbps package has a speed limit of 2 mbps, and the 10-megaps set is achieved by two means: port level limits and application level limits. The port speed is on AC or switchboard, with a bandwidth limit for users; the application speed is on the application gateway, with a different type of application. The costing effect is different: at the port level, the excess traffic is discarded and the billing system counts traffic after the speed limit; at the application level, the flow rate is broken down and the application rates are separately, if the charge system is charged to the aggregate, the application flow is aggregated. This aggregation logic may be misconfigured.
Drop package retransmit double-counting risk to traffic
WiFi network billing systems can be re-counted if they follow the original message flow. For example, users download a 100MB file because it is difficult to lose 10 MB packages and the actual transmissions are 110 MB, which may be recorded as 110 MB. The difference of 10MB is not significant for individual users, but the whole lesson is a very large number.——Whether to count as a message or as valid data. If the system does not support de-weighting, speed containment strategies need to be properly relaxed, reduce the probability of dropping packages and control duplicate traffic from its source.
Time-limited speed and time-time billing
WiFi in the school dormitories has a clear time feature——The speed limit is matched by the time-time limit, or there are problems of mismatch. For example, the speed limit strategy works at 8 p.m. but it changes to the peak rate only at 9 p.m. and the middle one hour user is limited to a low-peak rate, with no less difference in experience costs. Time-time strategy shifts need to be uniform, speed limits and costing cannot be changed simultaneously.
Matching of the particle size of a flow orthopaedic form to the measured particle size
The particle size of the flow orthopaedic is matched to the billing. If WiFi network billing system is based on user-level, the flow orthopaedic is made at user level.——The bandwidth of each user is independent and cannot be affected by the use of one user. If the system is charged at a regional level, flow integer can be done at a regional level.——The entire dormitory has a bandwidth pool, with internal users competing. Both types of user experiences are completely different. User-level limits are fairer but more complex, regional levels are simpler and peak periods are worse. Which option depends on the school’s network resources and management philosophy, but the post-counting particle size must be consistent with the speed limit, otherwise they will occur."Speed limit is at the user level, but the billing is at the regional level."This mess.
Differences in traffic and deductions under prepaid settings
WiFi network billing system, which is a prepaid fee model, is used by users after charging and usually using flow-button mechanisms——The user will have a bad feeling about the difference between the button and the button if it is not properly handled. For example, the withholding of 100 MB, when does the remaining 20 MB actually go back? Is this real-time return or next month? If it is returned in real time, users may see the balance jumps frequently during the connection and experience bad; if it is returned next month, users will feel that it is too late to return."The money's been withheld and not paid.". The withholding strategy is well configured: the amount of the withholding should not be too large (usually 5-10% of the package), return times should not be too low (one settlement per hour) and users must see a balance and actual usage that is generally consistent.
The flow calibration mechanism is the last guarantee of accuracy.
No matter how well a flow orthopaedic strategy is matched, it is impossible to be 100% accurate. WiFi network billing systems must have a traffic calibration mechanism as the last guarantee.——The system compares user traffic with the export gateway log, and the deviation is marked separately by users who exceed the threshold. Calibration mechanisms cannot eliminate bias, but they detect anomalies that prevent large area costing errors due to bugs or configurations. Calibration thresholds are generally set at 5-10%, while deviations below 5% are normal and more than 10% must be checked.