Italian Qo S Monitoring Network impact on User

Italian Qo. S Monitoring Network: impact on User Protection Regional Development Forum for Europe (RDF-EUR) Broadband for Sustainable Development 20 -22 April 2015, Bucharest Eng. Luca Rea Project leader for the Italian Resolution n. 244/08/CSP deployment

Outline • Project regulatory and technical aspects • Technical aspects: • End point location (Internet e. Xchange Point ) • Protocol type • Qo. S parameters and file size • Next Qo. S measurement metrics • Network architecture for ISP measurements • End users measurements • Ne. Me. Sys results • Ne. Me. Sys speed test Regional Development Forum for Europe (RDF-EUR) Bucharest, 20 -22 April 2015

Italian NRA (AGCOM) Resolution n. 244/08/CSP User Protection - Main aspects - Aims: • • ISP measurements: Number of measures to help consumers in understanding more on broadband service they are going to buy. Broadband results obtained by “ISP measurements” will be used by each Provider instead of “up to x Mbit/s” (advertised values) End users measurements: By means of the software Ne. Me. Sys (Network Measurement System) every end user can evaluate its own fixed line Quality of Service. Results could be compared with ISP’s promised ones Contents: • • • Measurement models Services supplied by operators Qo. S control and consumers’ rights Supervision by an independent body (Fondazione Ugo Bordoni) Sanctions Regional Development Forum for Europe (RDF-EUR) Bucharest, 20 -22 April 2015

Resolution n. 244/08/CSP Qo. S fixed-line broadband connection - Operators duties • Operators must publish the named “Carta dei servizi”: document containing the main principles that offered services must comply with. Minimum requirements in terms of Quality of Service are established according to ETSI ES 202 765 – 4 • Operators must transmit relevant data to AGCOM (Italian NRA), publish on their websites conditions of their services and advertise correct data as far as data transmission speed is concerned • Provide predefined contents for fixed-line contracts: ü communication indicating where information on aims and results concerning Qo. S can be found; ü communication concerning the technical aspects of the Qo. S provided, which must be available to the customer before signing the contract; ü technology used for the access (e. g. , Dial-up POTS, Dial-up ISDN, ADSL 2, ADSL 2+, fiber optic) and the minimum characteristics that customer’s CPE must have in order to guarantee the full and correct performance of Internet access contract; Regional Development Forum for Europe (RDF-EUR) Bucharest, 20 -22 April 2015

Resolution n. 244/08/CSP Qo. S fixed-line broadband connection – Consumer rights • Operators bind themselves to observe values indicated within the contracts, so if values measured are worse with respect to published ones, he can file a detailed complaint to the undertaking involved: • • • He can claim Qo. S level restored If proper levels of Qo. S are not restored within 30 days from the complaint, the customer has the right to recede without penalties from the contract. Each customer can formally certify the quality of his own fixed-line broadband access by means of the free software Ne. Me. Sys www. misurainternet. it Regional Development Forum for Europe (RDF-EUR) Bucharest, 20 -22 April 2015

Measure Architecture Measurement Server Operator’s IP Network Twisted pair Operator’s IP Network IXP Central System FUB Operator’s IP Network Twisted pair Management Consolle Client’s Location Acquisition System Measures processing Data Repository Both ISP Measurements and End-user measurements employ the same Network Measurements System, based on a PC software agent Twisted pair

IXP (Internet e. Xchange Point) IXP architecture One or more network switch, connecting the participating ISPs to each other. Providers 1 2 …. . n Server FTP 1 LAN Peering • It is a physical infrastructure through which ISPs exchange Internet traffic between their networks (autonomous systems). • The primary purpose of an IXP is to allow networks to interconnect directly, via the exchange, rather than through one or more 3 rd party networks FUB (ISP Measurements) 1 Gbit/s Router Up to 100 Mbit/s toward FTP 1 Up to 900 Mbit/s toward FTP 2 IXP Server FTP 2 LAN (End users ) Switch Through a measurement server located in IXP, it is guaranteed that each client measures only performances related to its own ISP Network. This approach has been chosen since it allows for Regional Development Forum for Europe (RDF-EUR) comparable and reproducible results Bucharest, 20 -22 April 2015

Qo. S parameters involved ETSI ES 202 765 -4 List of Qo. S parameters defined by ETSI Standard ES 202 765 -4 • Data transmission speed ü Data transmission rate achieved in a fixed time period (10 seconds) separately for downloading and uploading specified test files between a remote server and a user's computer • Packet Delay (one way transmission time) ü half the time in milliseconds, needed for an ICMP Echo Request/Reply Packet (Ping Command) to a valid IP address • Packet loss ratio ü the ratio between the number of not replied Ping commands to the total number of sent Ping commands • Unsuccessful data transmission ratio ü the ratio of unsuccessful data transmissions to the total number of data transmission attempts in a specified time period Regional Development Forum for Europe (RDF-EUR) Bucharest, 20 -22 April 2015

Qo. S parameters involved ETSI ES 202 765 -4 Regarding data transmission speed, the following four values relevant to data transmission speed are computed for both upload and download direction: • 95 th quantile in kbps • maximum bit rate during the analysis period of 10 seconds • average bit rate during the analysis period of 10 seconds The following two values are computed relative to delay: • • The mean value in milliseconds The standard deviation EG 202 057 -4 ES 202 765 -4 (up to 2014) (starting at 2015) Data speed rate protocol used FTP HTTP Throughput 20 times FTP (GET/PUT) FTP File size is equal to 10 time the headline speed under test. The results of each test is File length divided time to download We take into account number of Byte downloaded with a time slot of 10 seconds Delay 10 pings with a 10 s distance from each other (ICMP packet length is 32 byte) Regional Development Forum for Europe (RDF-EUR) Bucharest, 20 -22 April 2015

ISP measurements network architecture Network deployed on the whole Italian Country Bolzano BL TN PN UD Measurement Server: Neutral Acess Point (NAP) Trieste VR NO Milano Torino Genova Venezia PD MN PR Measurement Client: Ministry peripheral premises FO Bologna LU Ancona AR Firenze PI PG SI VT PE RI Sulmona CB LT Roma Secondary peripheral site (Dipendenza) Avezzano TR GR Main peripheral site (Ispettorato) FR BN AV CE SA Napoli Cagliari Bari PZ FUB central site: central system TA CS CZ Palermo MS TP CT R. Calabria Regional Development Forum for Europe (RDF-EUR) Bucharest, 20 -22 April 2015

ISP measurements (site characteristics) www. misurainternet. it/stats. php § Upload line attenuation is considered because it is less affected by frequency depending. § The Italian average distance between customers and exchange is about 1, 2 km and the corresponding average upload Line Attenuation is 11 db § 11 db in upload line attenuation & 1, 2 km between users and Central Office (Exchange) regards 50% of x. DSL consumers in Italy To have a comparison between ISP performances also Line Attenuation must be the same Regional Development Forum for Europe (RDF-EUR) Bucharest, 20 -22 April 2015

ISP www. misurainternet. it/stats. php Loop modem Ispettorato Regionale Cabinet PCs ”probe room” Twisted copper pair cable 0, 6 mm (22 AVG) or 0, 4 mm (26 AVG) Attenuation equal to 11 d. B in Upload Line termination Access Network OLO room Regional Development Forum for Europe (RDF-EUR) Bucharest, 20 -22 April 2015 Central office

ISP mesurements Example of Probe Room Regional Development Forum for Europe (RDF-EUR) Bucharest, 20 -22 April 2015

ISP mesurements www. misurainternet. it 5 th percentile, computed on the whole national samples, represents the value that each ISPs can advertise. In a short time the minimum will be 95 th quantile. All other values are available by local analysis at www. misurainternet. it Regional Development Forum for Europe (RDF-EUR) Bucharest, 20 -22 April 2015

End User measurements: Ne. Me. Sys (Network Measurement System) It is an open source software that allows to any consumer to certify himself his own fixed-line Qo. S • Free download from www. misurainternet. it ü Registration is needed • Cross-platform ü It runs under all major OSes : Windows, Mac OSX, and Linux • Before starting measures, Ne. Me. Sys checks CPE right environment ü Before of a single test user’s LAN is analyzed to assure that nobody share bandwidth it is going to measure. RAM occupation and CPU busy are also controlled. • Measures endure 24 hours ü To obtain a significant number of samples that allow to well characterize the fixed line under test, Ne. Me. Sys executes at least 50 tests per hour. 50 tests (20 download, 20 upload and 10 ping) constitute a measure. Regional Development Forum for Europe (RDF-EUR) Bucharest, 20 -22 April 2015

Headline Speed VS Actual Speed Up to now, more than 31000 users has completed fixed-line characterization Certified distribution The average Actual Speed is ever half with respect to headline one. In many cases, it is due to the distance Regional between user’s location and Development Forum for Europe (RDF-EUR) the central Bucharest, 20 -22 April 2015 office

Ne. Me. Sys Statistics Registered users User with at least one measure Emitted certificates In advance certificates 203. 938 58. 581 39. 757 21. 733 Emitted certificates Regional Development Forum for Europe (RDF-EUR) Bucharest, 20 -22 April 2015

Ne. Me. Sys Statistics Registered users Regional Development Forum for Europe (RDF-EUR) Bucharest, 20 -22 April 2015

Resolution n. 244/08/CSP Qo. S fixed-line broadband connection – Misura. Internet Speed Test- • To give to users an easier instrument to estimate Qo. S, a short version of Ne. Me. Sys has been realized. Such a version is named Misura. Internet Speed Test, and it allows users to have an immediately result. • Obviously results do not have an official value from “receding contract” point of view, anyway crowdsourcing approach allows an ex ante information to evaluate the single ISP performance according to coverage • Sample collected are one the pillar of national statistic with the aim of monitoring the growth and evolution of broadband Regional Development Forum for Europe (RDF-EUR) Bucharest, 20 -22 April 2015

Misura. Internet Speed Test Up to now, more than 399160 users has used Misura. Internet Speed Test Regional Development Forum for Europe (RDF-EUR) Bucharest, 20 -22 April 2015

Conclusions The Italian Experience Next Challenges: ØExtend the monitoring/evaluating system to mobile connections ØCreate a detailed National’ Actual Speed Map to give right information to users Ø Exactly identify coverage areas and suggest solution for broadband connectivity according to 2020 European Digital Agenda Target (i. e. LTE, LSA Carrier Grade, FTTx) Ø Stimulate competition between ISPs This map is obtained by Telecom Italian Wholesale portal. It is an approximations because of each local municipality perimeter has been colored according to the ISP presence and technology Regional Development Forum for Europe (RDF-EUR) Bucharest, 20 -22 April 2015

Italian Qo. S Monitoring Network: impact on User Protecion Next Meeting: Villa Griffone, Bologna – Italy The head office of Guglielmo Marconi Foundation is in Villa Griffone, the place where the Bolognese inventor began to develop the system of wireless telegraphy which then spread throughout the world. Villa Griffone is located in Pontecchio Marconi, about 15 km from Bologna, Italy. The garden and the rooms of the villa are home to the Marconi Museum which is dedicated to the origins and development of radio communications. Villa Griffone is a National Monument Thanks for your attention Luca Rea lrea@fub. it Regional Development Forum for Europe (RDF-EUR) Bucharest, 20 -22 April 2015