It was in 2005 that I started working on XtraWiFi. I must have bought dozens of different models of wireless routers, hacking them & try to make service like XtraWiFi. To be honest, the kind of clarity I have right now was not present at that time.
From that time onwards, every 1 1/2 to 2 years, I spent at least a month on this project and every time rejected the project
In 2012, I sensed that computing devices were available with people with low income group. It would be far easier to provide the service directly to a mobile phone. It is a different matter that same technology could be used to provide service to laptops, tabs and even desktops.
However, there were two major challenges even after hacking and taking complete control of the router.
- Computational power of the wireless router
- Non-availability of technology that could make the service offering possible on low cost wireless.
In 2014 I worked on various projects that could one day be sewn together for making such service offering.
One of the reasons why we can offer services at such low price is the fact that the majority of trafficking of torrents are not allowed in XtraWiFi and users cannot tether or use any other mode of sharing XtraWiFi services.
Why should a user cross-subsidize?
The torrent problem
One such project was related to torrent traffic on the Internet. I had access to lots of broadband companies' traffic. I analyzed the traffic and found a heuristic function to define torrent traffic. It could identify good torrent traffic and bad torrent traffic. I started developing tools to limit & block torrent traffic. It was extremely successful. I incorporated that into our bandwidth management solution, Netcontrol. This software is popular amongst broadband service providers in Hyderabad. I never intended to take Netcontrol bandwidth management solution to market in a big way to compete with similar solutions like Inventum. I decided to keep the technology low-key.
I challenge anyone to come with algorithm to identify Skype (which is a good p2p) from utorrent & other file sharing p2p (which are bad p2p. This is not meant to offend anyone).
In some of the networks, torrent traffic is up to 60%. People choke the network with torrent traffic. Any broadband service provider will have nothing but tears about this.
In continuation of efforts in this area, I developed a caching engine for torrent traffic. Christened xCache, this p2p/torrent caching solution has become popular in some parts of the country (West Bengal, Delhi, Mumbai).
But xCache was not the final solution I was looking for. Even this was kept in wraps for some time. I found one partner in Delhi to take the product to market, which is running smoothly independently.
The wireless router problem for broadband companies
When we started broadband services, they were terminated on a desktop computer. Slowly people started buying wireless routers to share the bandwidth to multiple number of devices. Currently 98% of all broadband connections terminate on a wireless router. The sharing ratio of bandwidth has plummeted. There were days when we used to share @ 1/30.
Those were the golden days of broadband service providers. As the cost of Internet bandwidth itself had started going down, the problem about the number of devices had not surfaced.
Broadband companies provide shared bandwidth to its subscribers and those subscribers share it internally; sometimes even with others. I could narrate cases like hostels and bachelor accommodations that bleed broadband companies.
These days, 10% of users consume 90% of bandwidth. If I could eliminate the top 10% of users, broadband service cost would come down and more low income group people can take these services. But it never happened in the industry. There was no technology available in the market to address the problem.
The only way to solve the problem is to devise plans with limited number of devices behind the router, which is a tough cookie to crack. Nobody has ever calculated the number of devices behind a router.
In 2015, I cracked it and netcontrol could now calculate the number of devices behind a router. Then there was the problem of computing power for calculation. Over a period of time, I optimized the algorithm for device count calculation. It is incorporated into netcontrol in 2016 for now and netcontrol can calculate the number of devices on the fly on a P4 device for 1000 subscribers with 6% average utilization. This is quite an achievement in the industry. I started offering limited devices service at a lower rate in one of the broadband companies I own.
One of the reasons why we can offer at such low price is the fact that the majority traffic of torrents are not allowed in the network and user cannot tether or share XtraWiFi services in any other way.
In the process, I developed netcontrol into a cloud based solution with minimal computing at the local server end; developed logs & backup systems to ensure business continuity. Cloud storage based daily & weekly backups are taken at two different cloud storage points with one old-school DVD burning as well.
We had back end infrastructure to not just handle the money/chips flow within the system but really optimize the processes.
Incorporating roaming into system was another major challenge. It is obvious that each administrative division would need different database system. Thus there would be multiple servers that would authenticate a user. I designed interconnecting mechanism that would make roaming a very smooth process.
An XtraWifi user in Hyderabad can go to Delhi and start roaming services. Currently there is no different pricing for roaming. If a situation demands, there exists a mechanism within XtraWiFi to charge the user for roaming.
Since there are no mechanisms that are available with low cost, I ended up designing every module in the system.
The challenge of low computational power of radio was solved by redesigning every module with CPU cycles in mind. Our current wireless router utilizes about 12% of CPU on the board.
The captive portal of XtraWiFi is smooth and powerful. The user registration process is simple: the user needs to enter their phone number and authenticate by entering OTP into captive portal.
One hour complimentary service is given @ 4mbps to the user on successful registration.
A user can
- Add money/chips into their account
- Renew service for oneself
- Renew service for others
- Transfer money/chips to another account
It was in second quarter of 2016 that I started putting the entire solution together.
Testing & Trials
In June of 2016, first completely integrated solution was made. I offered 30 day 4 mbps service on registration. My neighbors were thrilled to bits. 45 people used the service in June while integration was being done.
I took it out of labs and started implementing in King Koti network. It was difficult to explain to the team that they are just witnessing a paradigm shift in broadband industry.
The response from the market was overwhelming. I recovered my hardware cost within one month of installation in King Koti.
Now plans are afoot to launch the service commercially.
I insisted on having a plan for Rs. 10 recharge. Now, one can watch two movies for ten rupees on one's smart phone. This is my dream come true.
My stated vision is generally perceived to be the job of the government. I have realized with my little experience that any public utility service that can be run by a private organization while making money (not as charity) is more efficient than Government subsidized service. XtraWiFi is a "for profit" service that I hope to reach every corner of the country.
Government support will go a long way to make this service ubiquitous. The following support is expected from Government for XtraWiFi project.
- Permission to lay cable on electric, municipality poles
- Permission to install CP box in government buildings, including Village Panchayat office
- Permission to install Wireless radio on any government building or pole
- Power connection with meter wherever required. Central billing. Subsidised or domestic billing
- Protection from Cable mafia (who think that they control a geographical area and do not allow others to operate)
Support offered to the government:
Any content that is useful to a farmer or any other citizen, like the weather app or government service apps can be offered on XtraWiFi platform without charging the citizen. Such service will be available on XtraWiFi platform everywhere.
By early 2017, we will have our first rural broadband implementation that will make profit. I expect about 60% of rural market can be reached within existing pricing structure. With little bit connectivity support from the government, each and every village in the country could be running on XtraWiFi.
XtraWiFi will generate numerous jobs in rural and semi urban markets in the days to come.
XtraWiFi is in tune with my life's vision "to positively contribute to the quality of the lives of people".
My vision for XtraWiFi is to provide high speed & inexpensive broadband to everyone.
I hope to bring inexpensive & super fast services to every corner of the country & play my role in contributing to the building of my country.
Happy Independence day!
15th August 2016