Il Bandometro isposure: question time con Ben Wood, responsabile di Epitiro per l’Italia
A soli pochi giorni dal lancio ufficiale dell’Italian Broadband Quality Index (iniziativa di Between ed Epitiro per l’analisi e il monitoraggio delle prestazioni della banda larga in Italia) abbiamo ricevuto dalla Rete moltissimi commenti positivi, ma anche critiche e richieste di informazioni: citiamo tra tutti i blog di Alessandro Longo (che ha pubblicato anche un articolo su Nòva Sole 24 Ore) e Luca de Biase, e diverse discussioni sui Social Media: http://tinyurl.com/yzbbuyg, http://tinyurl.com/yl2txs7. Questi feedback confermano l’importanza e l’attualità del tema della qualità della banda larga, che, a seguito della delibera AGCOM 244/08, sta prendendo sempre più piede anche in Italia.
La risposta alla maggior parte delle domande emerse dai blog e dai Social Network si desume direttamente dalla sezione FAQ del sito ufficiale di isposure. Tuttavia, per avere un riscontro diretto e puntuale sulle questioni sollevate, abbiamo intervistato Ben Wood, responsabile di Epitiro Technologies per l’Italia.
Epitiro is an International company that launched isposure in many countries: how did you answer the remarks from the users in New Zealand?
The thread under discussion is from a New Zealand forum and goes back to 2008. A few users claim that isposure results can be different from speedtest.net and that sometimes there are problems in using the software and the website. The issue with the speed test results was due to the huge popularity of Isposure beyond our expectations (we expected 2-3000 and received over 7000), we upgraded the capacity and as you will see from later forum posts this issue was resolved.
At the end of the thread, Mike Cranna, responsible for Epitiro in New Zealand, explains that “we can’t comment on individual test results” because “there could be a range of reasons why your performance is as it is”. For example, server location or testing methodology influence the results. In Italy the speed test is a TCP test run with multiple sessions until we maximize out the line capacity, to an endpoint server in Amsterdam, other speedtests often report on the maximum speed achieved over a single session to a local server. All the European isposure agents use the server located in Amsterdam, in order to guarantee the same conditions for the international benchmark. In the next phase of the project with Between we will move to a test server based in Italy.
Regarding bugs, software is rarely perfect on the first release. Following user feedback and collaboration with local partners we work daily to fix the bugs and improve the features. In Italy, for example, before launching isposure we did three months of testing. Please note that in New Zealand Isposure was launched together with the NZ commerce commission, the telecommunications regulator and they are still using the data produced by Isposure to monitor the performance of the ISPs.
In the UK Epitiro works in partnership with Thinkbroadband.com, a consumer association. There was a claim made on the forum that isposure operates similar to a spyware.
This was simply a problem of communication: isposure is included as part of the download for the Thinkbroadband tbbMeter software, however this was not communicated clearly enough in the beginning. It was not made clear that the data is collected and provided to ISPs. This was only an issue as the users were not downloading from Isposure.com, where we clearly communicate that the data is collected and may be supplied to ISPs in order for them to analyse the performance of their network and make inprovements where needed (for example adding capacity). Transparency is very important to us.
In Italy we have made it clear on the isposure website that the software is a piece of freeware, but also that we may be funded by the ISPs in future because we need to cover the costs of development, maintenance, upgrades and bandwidth for the end-point server. We do not collect personal data about Internet usage and it could well be in the interest of the users for the ISPs to see performance data in order to allocate resources to areas where the performance is bad. All users are informed of this and have the option to download or delete the software.
Ping time: why the websites used for the Ping test are representative of online gaming user experience?
Online gaming requires a good international ping performance, when an online gamer joins a game he will be entering an environment populated by gamers from locations all over the world who they will begin to interact with. A good ping performance is important in order that all the gamers are synchronized regardless of where they are located.
Gaming servers are private and not open to external applications to ping, so we use webservers based around the world (we continually verify their location) in order to run the ping test and provide a ping performance from the end-user connection around the world.
In addition to ping times and http/tcp speeds, isposure tests also the DNS lookup time: which DNS are used and why the testing website are representative of DNS performance?
We use actual website servers. This is the same as with all our tests (except TCP speed) as this is the only way to replicate the user experience. A normal internet user will visit websites which are hosted all over the world, if you want to understand DNS performance from the user perspective you need to understand how well the ISP network is handling DNS requests to normal websites: this is why we use 4 Italian and 5 international websites at key locations around the world.
So, which criteria do you use to select the test-websites?
Regarding the test locations, for all tests we use two main criteria. The first one is location, as I said before we test both Italian and international websites to reflect the normal usage of the end-user. For example, for HTTP test we use 4 Italian and 5 international locations, while for ping test we use 4 Italian and 14 international sites to represent the much more international nature of online gaming.
The second criteria are the website attributes, we make sure that the website has enough rich content for a valid test (HTTP) and good underlying performance (for all the tests) to ensure that the performance of the website does not affect the results.
Today it is very easy to find freeware for the analysis of the Internet connection, and Isposure has to face a strong competition. Why is Isposure different?
The performance of your Internet connection could be affected by many variables, such as distance from the server, distance from the Exchange, time of the day, peer-to-peer activities etc. A single measurement of your line speed does not provide a comprehensive view of the performance of a broadband connection. In order to understand the real performance of your Internet connection is important to look at the average results of a longer period and to use a set of metrics and not just TCP speed or Ping times. Isposure is different because:
- Continually monitors the performances over time
- Allows users to register many profiles (one per each different connection)
- Other than TCP speed, Isposure also measures HTTP download speed, Ping time and DNS lookup time
- Allows users to compare their results with the use experience of other customers in the same area
- Allows users to easily check the results of all the tests, displaying the averages of the last weeks
Isposure is a software that analyses my connection, collects and send data to a server: many users are worried by the privacy issues.
In the FAQ page of the Isposure website, we explain our privacy policy clarifying which data is collected. We respect our customer’s privacy over all else so we never monitor, collect or transmit information about the web site visited, any 3rd party software or application loaded, Internet activity or any other Internet-related action performed. We collect just the data about the performance of the Internet connection.
Finally, I would like to specify that many of the issues pointed out on the blogs or on the Internet discussion groups are now fixed by the new releases of the software. Anyway, for problems, doubts or information please visit the FAQ page of Isposure website.
L’8 ottobre a Capri, in occasione del convegno “
