Pavel Durov

In our case, we have tens of thousands of servers around the world, almost a hundred thousand, distributed across several continents and data centers.

If you try to manage this system

manually without automation, you will probably end up hiring thousands of people, tens of thousands of people.

But if you rely on algorithms and the team is forced to put together algorithms

in order to manage it, then it becomes much more scalable, much more efficient, and interestingly, much more reliable as well.

Yes, humans are attack factors.

And if you have a distributed system that runs itself automatically, you have a chance at increasing the security of speed and speed of your service.

This is what we did with Telegram.

while also making it much more reliable.

Because if some part of the network goes down, you can still switch to the other parts of it.

The way we designed Telegram is we never wanted to have any humans, any employees have any access to private messaging data.

That's why since 2012, when we've been trying to come up with this design, we always invested a lot of effort into making sure that nobody can mess with it.

Like if you hire an employee or any of the existing employee, they can't break the system in the way that would allow them to access messages of users.

And then, of course, we launched into an encrypted messaging that is even more protected, but it has certain limitations.

So you still have to rely on encrypted cloud.

So an interesting engineering challenge was how you make sure that no point of failure can be created within your team or outside.

No, never.

Telegram has never shared a single private message with anyone, including governments and intelligence services.

If you try to access any server in any of the data center locations, it's all encrypted.

You can extract all the hard drives and analyze it, but you won't get anything.