Three years with Turris

A couple of weeks ago, I got an email informing me that it had been almost three years since my entry into the Turris project, and I could now purchase the router for a symbolic price of one crown. I did that right away to test for my colleagues whether the system works well; however, it also brought back nostalgic memories. Because three years ago I had the same goal – to test whether everything was working properly – when I filled what was probably the first router lease contract. Those three years have gone by in a flash, so it is perhaps a good time to stop and look back.

If I was to briefly summarize the period, I’d say that we have come a long way. We have moved from the first thousand blue Turris devices intended primarily for research purposes into a situation where we have just finished one of the most successful Czech crowdfunding campaigns (the world’s 116th highest funded crowdfunding project), we are offering our routers commercially and we meet with such interest that we are already preparing the next batch. However, the way to this success certainly was not easy or straightforward.

The preparation of each new batch of routers took us at least a year. Paradoxically, the fastest was the preparation of the first version, which went to the production 12 months since the launch of the project, even though for the first three months we had no hardware developers and we had no idea how such things are made. It was a time of great enthusiasm, lots of overtime and stress, but also of fast progress. The second version, where we improved the power supply and cooling and, most importantly, added USB 3.0, took us a little longer, even though the changes were not large. This was largely due to the fact that the team already had a lot of work with the maintenance of existing Turris devices, to the detriment of time for innovation. Our third endeavor was Turris Omnia. Despite the change of platform including processor architecture, we have succeeded to quickly produce functional prototypes with which we entered a crowdfunding campaign. If you want to know how much went wrong afterwards in later stages of development, you can listen to the excellent presentation by Ondřej Filip. However, within a year since the start of the campaign we have sent out all the promised routers, which is still a decent result compared to average hardware crowdfunding campaigns ;-).

Not to dwell on hardware only, software development was of course also interesting. For the Turris router we developed an automated upgrade system, which met with positive response at last year’s OpenWrt Summit, among others. It is true that not everything was done without mistakes, but I still believe that we went about it the right way. By the way, Turris 1.0 still uses the recovery system from 2014, which is launched in case of a factory reset and which is able to update the device to the current version of the system, including the exchange of the libc library, significant kernel upgrade, and so on. It is obvious that we had the right idea that time.

And since I mentioned the acclaim at the OpenWrt Summit… The fact that we have become a full member of the community of this operating system is evidenced by the fact that we managed to push through Prague as the city of this year’s annual OpenWrt Summit. And because we are going to be its host, I believe it will become the best one so far.

But getting back to the software, there is really a lot to mention. Our website contains a wide range of public statistics and information for users of individual routers. Recently, we have added the service that can help you find out if you are infected with malware. By the way, it uses also the data from honeypots active on Turris routers, which last year enabled us to be the first in the world to report the activity of the Mirai botnet (its name was still unknown at the time), comprised of online cameras and other smart devices.

And that brings us to another area of development, namely the network security. Many results of our work are visible, but those that are probably the most valuable are hidden from the public. These include cases when various kinds of abnormal behavior of network devices led to discovery of a malware in some turrists’ network. In such cases, in cooperation with the CSIRT.CZ team, we discreetly established contact with the user and often managed to find and eliminate the infection together. There have been several dozens such cases in the past three years, and even though it is not a staggering number, I personally really appreciate each and every one of the cases, when we could help the user.

So when I look back at the plans we had when starting the project, and zoom out a little from operational things and details, it’s hard for me to find a reason for discontent. We fulfilled our ideas in all key areas, while in 2013 we could only dream of the global success in sales of an open hardware “Made in the Czech Republic”.

As I mentioned, the path of the project Turris from 2013 until today has not been easy. And it certainly would not have been possible without the right people. I would like to conclude this brief lookback by thanking everyone who was and still is involved in the project Turris: the developers who have been coping with the strenuous workload for years, the CZ.NIC management that gave us confidence and did not hesitate to invest considerable resources in the project, and certainly our users and fans — turrists, who help us with their feedback and have patience with us when something goes wrong.

May the next three years of the project Turris be at least equally successful!


Zanechte komentář

Všechny údaje jsou povinné. E-mail nebude zobrazen.