Post nostalgia: how we started Acronis at MIPT and in the attic of the upper-air observatory

2006, Dolgoprudny. Celebrate someone's birthday. Alexander Chernykh, Alexander Golubov, Valery Belous, Dmitry Chepel

For me, the company started with four people. This is me with my brother Oleg, Yura Per and the leader Maxim Tsyplyaev. Before that, we worked together in Fiztech-software under his leadership. Our then-product was called PTS BootWizard, then it became Acronis BootWizard, and then Acronis OS Selector. In principle, it was a product that included a basic set of technologies that could be developed further.

Since 2000, we have worked in a new company called SWSoft, which makes ASP Linux. Actually, the technology was bought out from the owner of “Fiztech-software” (including the resizer and BootWizard) and the team.
')
I remember at the very beginning the computers we had were very cool at the time: two hard drives of 20 GB each and CRT monitors of 17 inches each. Heavy, we often wore them because of travel. They wrote in C ++ on Watcom, part of the code was assembled under DOS. At the very beginning, there was no code for Windows at all.

The difference between “Fiztech-soft” and SWsoft was obvious in such details, for example, as feeding. In “Fichtech-software” we ate at our own expense and as it should. But in SWsoft everything was already clear and conveniently organized: there were aunts who cooked food, brought it. Delicious, such homemade food. The company paid.

Let's go back a little to ASP Linux - in fact, SWSoft needed a resizer for the installer, looking for a team that could perform this task. So come on us. Resize we knew how.

SWSoft at that time had already opened an office at Fiztekh, and we just moved a couple of floors through our friendly development. We sat in a separate room - in the laboratory building of MIPT. Then the product line on the basis of the written code then began to be sold, and SWSoft singled out the future Acronis into a separate company. So we got there.

A new name appeared in January 2001 (at the same time the domains COM and RU were registered), I wanted alphabetically to get up the sorting order - therefore, “AC”. There is no special semantic load: from the dictionary we wrote out a few favorite words, then somehow we came up with it. Officially, the company appeared in 2003. At the same time they decided on the management: Sergey Worked directly with Max, Max was de facto in charge of everything then. Stas Protasov acted as an external consultant. Today Stas leads the entire development, Sergey is the CEO, Max Tsyplyaev and Ilya Zubarev are co-founders.

Almost immediately in 2001 they took Dima Boytsov to the site, Lesch Katkov came. Pretty soon it became necessary to move to another building, did not fit at all. Max Lyadvinsky came in and began to make snepshots technology in order to back up a lively Windows at the sector level. He previously worked on drivers and kernel stuff. He wrote the library at home in the evenings and looked for a long time to apply. In Windows then there were no own snapshots, and other vendors of software for backup also did not have them. From competitors then were Norton, Veritas, Poweerkvest. We walked all of them on a low level and made the first snapshots.

Yura Per was just doing what he did a quick resize. It was necessary for the minimum time to turn one section into another - located differently and resized. Minimized the number of copies of the data: if the source and destination regions overlap, then if possible, try not to touch the data. In Windows then there was no tool of his own, only later he appeared rather clumsy (only in a big way and with restrictions), and then he practically did not develop.

FAT16 had different cluster sizes: the code was able to change arbitrarily up or down, respectively, it was possible to increase the FAT16 partition to wild volumes and simultaneously convert it to FAT32. For FAT, they simply wrote to the first hole: a displacement map was compiled, first virtually, and then the displacement itself went, and in the process of moving blocks with data (for example, when the directory was shifted), the cluster numbers were immediately changed to new ones. The new allocation table was in RAM, so the data was mixed during an emergency power outage. It was important not to touch the switch in the process.

To change the beginning of the section downward quickly, especially nobody could, this is a more difficult task. By the way, it was ASP Linux in 2000 for its installer, the first one who knows how to correctly “push” Windows and make its partition “on the fly”, worked just with our code - it was this code that I wrote, and it was for this reason that it was purchased together with the team.

It was difficult with the graphic library: you had to either write yourself (but do the window library for a long time), or pick up: QT was already (but young and flawed, without full-screen options), Zinc did not develop, Fox remained.

Oleg parallel riveted products. Then the term for the product, taking into account the already available low-level base, was absolutely fantastic by today's standards - for something it was 2 weeks, for example. We needed a ruler and quickly. Now 50 people can make a product for several months, and then one person sat down and wrote for two weeks, then another documents, and can be sold.

For example, the first release of Acronis Disk Director (which was then called Acronis Partition Expert) was a clean GUI to our DOS libraries for working with the disk, and everything else was written. Then we started porting the code under Win. Or, for example, Acronis True Image, resizers. It was necessary to do just a step-by-step wizard, an interface that gave input commands to the library. The idea was simple, all the components for implementation were already needed, only the GUI was needed. Exaggerating, if desired, it was possible to write “on the knee” in 1 day, but until that moment, none of the competitors had thought of and did not write. This is how we were the first to make backup simple and easy to restore. After the product became popular, we repeatedly completed it, improving it today, but the first version was assembled quickly, in about 2-3 weeks completely.

The tests were then done directly on the developers' machines: they themselves did the testing, and they themselves did the support - they wrote the manual, and the builds of the binaries themselves too. Now all this is done by some people, and then it took me 20 minutes a day to answer questions by mail from Germany, Japan and the States.

Yura and I were doing low-level disk technology. Read-write, backup, restore. Jura is a real guru of file systems and their support. I was digging with disks on an almost physical level: partition structure, read-write, deferred execution of operations. The same interaction with the snapshot driver. There was support for FAT, then NTFS, ext2, ext3. There was also support for Sun, IBM systems.

Version control was done via CVS: you pump it out in the morning, and you pump it out in the evening, then commit it. Then, when there were more people, it became more difficult with it: the settings were not the most obvious, for work it was necessary to enter a terrible line. Although it was under Windows, but her legs grew from * nix. There was a CVS with us for quite a while. At the very beginning, we had one Linux server that simultaneously served as a gateway, mail server, file server, and CVS. SVN came late when there were more people and code. Around the year 2006–2007. Then there was no review - anyone could write anything, and it went into release without checking by other participants. Although the cap, of course, from each other on the tests ogrebali. Unit tests were not, there was only author testing. Easter eggs were not, somehow it was not accepted. QA in the modern sense was not, Denis Dementiev was responsible for both bugs and support. Later, in 2002, a dedicated QA appeared - Valera Drozdov from the Falcon came, he was in charge of them. We still "rented" Vladimir Simonov, and then he stayed with us. Until now, one of the key developers for Mac and * nix.

Bugs have always been enough, nothing is perfect: we are constantly working with different hardware, different BIOS, with a low level. There was a new version of Windows, I had to sharpen under its features. The BIOS was then updated from a floppy disk, and everyone had different versions. It was often that you are debugging the program in the same place where you write on the same machine. You set up, launched, you start to reduce or increase the section - and that's all, you cover it. Virtualization then really was not.

I must say that the year 2002, in terms of backup, was already quite developed. This is not "the most important thing on diskettes and in the safe", there were already different branches of technology. Most often, backup data was saved locally to the hard disk, for example, to the second physical disk in the system unit or to its hidden partition. There was support for CD and DVD, and later tapes were needed when many corporate clients arrived. By the way, from the development point of view, the corporate backup was not so interesting, there are thousands of additions for specific requirements. And the home backup was space, there were constantly made important functions that everyone needs, and each new iteration was noticeably better than the previous one. It is very inspiring.

There were boot floppies with recovery and boot zip floppies. Our product offered to save everything, and then burn the disk for emergency boot. If the user only had a 1.44 drive, he needed 4 floppy disks, and they were inserted one at a time during the boot. The CD went later, and then the main sales were in the form of boxes, and there was already a CD in our box from which it was possible to boot to recover. Then the Russian craftsmen made many such things out of him:



Later, with the release of "whists", a lot of technology has changed. Then we rewrote almost the entire stack, made Try & Decide, when you could enter a special mode and commit all changes, and then roll them back if you don’t like them. It was cool to try a new virus. I remember Oleg just for Try & Decide did the bootmenu, which offers to roll back or apply before loading the Windows.

Offices have changed. First, from MIPT we moved to the Central Aerological Observatory: we had the entire sixth attic floor. It was comfortable, spacious: a lot of rooms are empty, no one was sitting at first. Then people scored, filled. I remember that we were driven out of alarm by everyone, the exercises went regularly and hard. By the way, before us there were chemists some, it seems, or pharmacists, I do not remember exactly. There were signs on the doors. We then called each other part of the office as a joke.

With the Internet it was hard: then the traffic was limited. Everyone knew who shook and how much. Then, later, it was difficult to get used to the fact that new people swung whole (!) Hard drives and considered this the norm.




2007 - a strong development team gathered. Max Ammosov, Valery Belous, Dmitry Chepel, Grigory Zubankov, Nikolay Popov, Alexander Fomichev, Anna Tyuryumova, Alexey Kosarchuk

Then in 2007 we moved to the “Savyolovskaya”. There was a huge hangar about 50x50 meters, and our room was hand-made literally under the ceiling of a huge complex. Down there was either a market or a shopping center, and we were laid on the floor, and a floor without windows was formed. During this period, we have a noticeable division into managers and developers. Why it is noticeable - we just bought new screens, and LCDs began to issue managers, and we still had CRTs. On the LCD monitor it was possible to determine that the person in the company is new. Although the first LCDs we set back in Dolgoprudny, the gradation was noticeable. At the same time, the whole line was formed and the stack of technologies was more or less settled.

Then we worked very hard on an important product - Acronis Backup & Recovery (then it was called Acronis Echo Enterprise Server). Then there were two offices, we moved to our current on Altufevskoe highway. Yura and I were on duty each in turn - it was necessary that there was an expert everywhere. By days it was distributed — Monday, Wednesday, Friday — I am there, he is here.

And then they moved everything, and modern history began.