Franz Seiser

Same for data.

Typically at nighttime, you may not consume that much.

So if you disconnect you for a few seconds in nighttime and bring you back, it's a different IP address you don't recognize.

This switching over can be done relatively easily, but it needs tremendous amount of planning, but that's a manageable problem.

It doesn't help to avoid downtime because Telcos always took pride in having this famous five nines availability.

We also did it in the old world.

The new system, though, helps people like me who are in charge for certain part of the systems to sleep a little bit better.

Because you have more resilience in the network, right?

There's always something going on and we do tremendous number of updates and changes in the network all the time.

And if in a classical setup, you only have two or three sites and one of the sites is going down, you're suddenly having two or maybe even only one lag.

In a cloud environment, that's fundamentally different.

Let's take a simple example.

You have one network function.

You need 20 instances to deal with the capacity.

If you have all 20 running on one location or you have on five locations, always four,

doesn't make a difference.

The cloud is just located at different places.

These four instances are running anyhow independent of each other.

So if I then lose one of the five, I still have four left.

And if you go to something like 10 sites, it's even better.