In case of any doubt regarding interpretation of any of the provisions of these regulations, the clarification issued by the Authority shall be final and binding.
(Rajeev Agrawal) Secretary Note – The Explanatory Memorandum explains the objects and reasons including measurement methodology for various Quality of Service parameters of the “Standards of Quality of Service for Wireless Data Services, Regulations, 2012 (26 of 2012)”.
Explanatory Memorandum
1. BACKGROUND:
1.1 Section 11 (1) (b) (v) of the TRAI Act 1997 mandates the Telecom Regulatory Authority (TRAI) to “lay down the standards of quality of service to be provided by the service providers and ensure the quality of service and conduct the periodical survey of such service provided by the service providers so as to protect interest of the consumers of telecommunication services”. In the discharge of these functions and in order to,…….
(i) create conditions for customer satisfaction by making known the quality of service which the service provider is required to provide and the user has a right to expect;
(ii) measure the Quality of Service provided by the Service Providers from time to time and to compare them with the norms so as to assess the level of performance; and
(iii) to generally protect the interests of consumers of telecommunication services, the Authority, in exercise of its functions under the above provisions in the TRAI Act, had notified the “Quality of Service of Broadband Service Regulations, 2006” on 6th of October, 2006. These regulations provide benchmarks for the Quality of Service parameters of wireline broadband services. TRAI has also laid down the Quality of Service standards for dialup and leased line internet access through the Regulation on Quality of Service of Dial-Up and Leased Line Internet Access Service, 2001 dated 10th December, 2001.
1.2 With the roll out of 3G and Broadband Wireless Access (BWA) services, the growth rate in cellular wireless telephone service is poised for higher growth, compared to wireline internet users. Presently, there is no quality of service standards for the wireless data services. It is necessary to benchmark and monitor the quality of service for wireless data services offered by the service providers so that the interests of consumers are protected. The Authority, therefore, decided to benchmark the quality of service parameters for wireless data services.
1.3 The draft Standards of Quality of Service for Mobile Data Services Regulations, 2012 were released on 09.07.2012, seeking comments of the stakeholders. Subsequently, an Open House Discussion was also held with the stakeholders on 10.10.2012. Keeping in view the comments received from the stakeholders, these regulations have been formulated.
2. Measurement setup
2.1 Specifications:
2.1.1 The measurement setup to be used to conduct test calls for measuring successful data transmission download and upload attempts, minimum download speed, average throughput and latency is given in Figure 2.1. The basic measurement set-up consists of a Test-Device and a Test-Server with specified software and hardware. Test calls have to be established between the Test-Device and Test-Server and measurements must be made for the respective QoS parameters. These parameters shall be measured in a stationary mode. Service Activation/Provisioning, PDP Context Activation Success Rate and Drop rate are to be reported from the actual network counters/database.
2.1.2 To assess the quality of the connection between an end user and an Internet Service Provider (ISP), ideally the Test-Server should be placed as near as possible to the gateway providing the interconnection between access network and ISP network. The location of the test-server as near as possible to the gateway providing the interconnection between access network and ISP network implies that the measurements will not reflect the influence in the QoS of the ISP network, between that gateway and the gateway interconnecting with the Internet.
Figure 2.1: Measurement set-up
2.2 Requirements for the Test-Server:
2.2.1 For all tests, a dedicated test server should be used as a well defined reference. The test server may be located centrally for all the licensed service areas (LSA) or for a number of LSAs or in each LSA (not more than one in each LSA). Under no circumstances a commercial server (e.g.
www.yahoo.com) should be used, since the test conditions for such a server may change over time making later reproduction of the results impossible.
The test server should be identified by an IP address and not by its Fully Qualified Domain Name (FQDN) in order to avoid issues with Domain Name Server (DNS) lookup and including the DNS caching strategies of the used operating system into the measurement.
2.2.2. The Transmission Control Protocol (TCP) settings of the server tested against should also be recorded. Since the number of host operating systems for internet servers is larger than on the client side, no detailed recommendation concerning the TCP settings of the server is given.
Test- device (End user) Mobile access ISP However, the TCP stack of the reference server should at least be capable of the following:
• Maximum Segment Size between 1380 Bytes and 1460 Bytes.
• TCP RX Window Size > 4096 Bytes • SACK (Selective Acknowledgement) enabled.
• TCP Fast Retransmit.
• TCP Fast Recovery enabled.
• Delayed ACK enabled (200 ms).
2.3 Test File:
The test file should consist of incompressible data i.e. a data file that is already compressed, e.g. like a zip or jpg file. The test file should have at least twice the size (in Kbit) of the theoretically maximum data transmission rate per second (in Kbit/s) of the Internet access under consideration.
2.4 Representativeness or Number of test calls:
2.4.1 The choice of adequate test calls, i.e. geographical locations of origin and destination of calls as well as traffic variations, is a crucial point with respect to the comparability and validation of the statistics to be calculated for the measured parameters. For each parameter, it is to be ensured that the samples are aggregated over all classes of customers for fairness in reflecting the QoS actually perceived by the user and the statistics shall be preserved to substantiate the same.
2.4.2 The necessary number of samples (test calls) shall be 1067 for each of the category “A” and “Metro” licensed service area (LSA), 600 for each of the category “B” LSA and 384 for each of the category “C” LSA for all the parameters.
3. Frequency of measurements:
3.1 For measurement of QoS parameters, it will be suitable to take measurements on a monthly basis during Time Consistent Busy Hour (TCBH). The number of test probes (Test PCs) used in the area of coverage of the system of measurements should be enough to guarantee that the measurements are representative of the coverage area and sufficient from a statistical perspective. In order to guarantee the spatial representation of the measurements, the operators should establish and deploy test probes depending on the number of active customers in each city/town/region, which should take into account the market size and demographics in the license service area concerned.
3.2 For example, in a category A LSA, there has to be atleast 1067 test calls for each generation of the technologies. These test calls have to be spread over the LSA covering all the geographical areas (cities/towns) in proportion of the total number of active customers and shall be further proportioned to cover all the plans working in the LSA.
3.3 In some cases disasters, freak weather, etc. may distort measured QoS figures. Such occurrences may not necessarily damage a network, but could degrade QoS by inducing exceptional traffic levels etc. In these cases, service providers should provide the measured QoS and may additionally provide a second figure which excludes the effects of the exceptional circumstances. A note clearly explaining the difference should also be provided.
4. The QoS parameters:
4.1 Service activation/ provisioning:
This refers to the activation of services after activation of the SIM.
This involves programming the various databases with the customer's information and any gateways to standard Internet chat or mail services or any data services. The service provider typically sends these settings to the subscriber's handset using SMS or WAP.
4.2 Successful data transmission download attempts:
4.2.1 The successful data download attempts is defined as the ratio of successful data downloads to the total number of data download attempts in a specified time period. A data transmission is successful if a test file is downloaded completely and with no errors.
4.2.2 Measurement: The percentage that is the sum total of successful data downloads, divided by the sum total of all attempts to download a test file should be provided. The statistics should be calculated from test calls made according to the measurement set-up and taking into account the representativeness requirements. The successful data download is measured by downloading a test file. An attempt to transmit the test file should be considered unsuccessful if it takes longer than 60 seconds.
4.2.2.1 The threshold of 60 seconds refers to the limit for acceptable performance for bulk data transmission/retrieval of ITU-T Recommendation G.1010.
4.2.2.2 The transmission time is the time period starting when the access network has received the necessary information to start the transmission and ending when the last bit of the test file has been received.
4.3 Successful data transmission upload attempts:
4.3.1 The successful data upload attempts is defined as the ratio of successful data uploads to the total number of data upload attempts in a specified time period. A data upload is successful if a test file is uploaded completely and with no errors.
4.3.2 Measurement: The percentage that is the sum total of successful data uploads, divided by the sum total of all attempts to upload a test file should be provided. The statistics should be calculated from test calls made according to the measurement set-up and taking into account the representativeness requirements. The successful data upload is measured by uploading a test file. An attempt to transmit the test file should be considered unsuccessful if it takes longer than 60 seconds.
4.4 Minimum download speed:
4.4.1 The download speed is defined as the data transmission rate that is achieved for downloading a test file from a test server to a test device.
4.4.2 Measurement: The minimum download speed should be calculated from test calls made according to the measurement set-up. Test calls are to be made to weigh the results according to the patterns of real traffic.
Minimum download speed shall be average of the lower 10% of all such test calls.
4.5 Average Throughput for Packet data:
4.5.1 It is defined as the rate at which packets are transmitted in a network. In a mobile network the download speed varies depending on the number of users in a particular location. Even though a service provider may be advertising certain speed, the actual speed may vary as per the number of users in the network and there could be customer dissatisfaction on account of relatively slow speed. Hence, there is a need to prescribe an average throughput to protect the interest of consumers. The service providers need to constantly upgrade their network to meet average throughput benchmark.
4.5.2 The throughput is defined as the data transmission rate that is achieved for downloading a test file from a test server to a test device.
4.5.3 Measurement: The average throughput for packet data should be calculated from all the test calls made according to the measurement setup. Test calls are to be made to weigh the results according to the patterns of real traffic. Average throughput shall be the average of all such test calls.
4.5.4 The service provider will advertise the throughput being offered to its customers as per their category or plan and it should be meted out as per their commitment.
4.6 Latency:
4.6.1 Latency is the amount of time taken by a packet to reach the receiving endpoint after being transmitted from the sending point. This time period is termed the "end-to-end delay" occurring along the transmission path. Latency generally refers to network conditions, such as congestion, that may affect the overall time required for transit.
4.6.2 Measurement: Latency shall be measured with the test server for ping connected directly to the server on the same Intranet domain.
4.7 PDP Context Activation Success Rate:
4.7.1 A Packet Data Protocol (PDP) context specifies access to an external packet-switching network. The data associated with the PDP context contains information such as the type of packet-switching network, the Mobile Station PDP (MS PDP) address that is the IP address, the reference of Gateway GPRS Support Node (GGSN), and the requested QoS. A PDP context is handled by the MS, Serving GPRS Support Node (SGSN) and GGSN and is identified by a mobile's PDP address within these entities.
Several PDP contexts can be activated at the same time within a given MS.
4.7.2 Measurement: This measurement provides the number of successfully completed PDP context activations. For these context activations, the GGSN is updated successfully and a report of PDP context activation success is generated at GGSN.
PDP Context Activation Success Rate(%) = number of successfully completed PDP context activations Total attempts of context activation x100
4.8 Drop Rate:
4.8.1 It measures the inability of Network to maintain a connection and is defined as the ratio of abnormal disconnects w.r.t. all disconnects (both normal and abnormal). An abnormal disconnect may happen because of Radio Link Failures, Uplink (UL) or Downlink (DL) interference, bad coverage, unsuccessful handovers or any other reason. The drop rate is to be measured for all generations of the technologies separately.