Jeff Steiner

Like if we go back 10 years, at least in my area, in the database area, we have a pretty good idea what the next three, four years would bring us.

And I'm just not seeing that now.

I don't know what we'll be seeing in 12 to 18 months time.

And you can see that Oracle is dealing with the same things.

They've got a hardware division where they can put stuff on premises, but they also see that everyone's moving to the cloud.

So what do you do?

You can't just stop making hardware.

You can't focus 100% on the cloud.

Where do you strike that balance?

Well, I think it kind of depends what you mean by cloud.

customers just sort of up and moving the infrastructures that they own as they are now into the cloud isn't going to work very well for databases because the the data sets and the io requirements just exceed what you can reasonably do in cloud i did a big session at insight last year where i i fell all over myself trying to talk about how great i think amazon aws is they should have paid me to do that session um

But it isn't for everything.

You can't do everything that you've ever wanted to do.

And I barely answer the phone for anything less than 50 terabytes these days.

And just try it.

Try putting a 50-terabyte database that needs 200,000 IOPS in AWS.

Try that in Azure.

you can't do it i mean i guess you could but what you would have to spend to make that kind of performance real is just not going to make it worth it in my opinion that's also where nps happens to be coming in and playing a very nice role that's a net private storage where that overcomes a lot of these problems and i'm really not trying to make this a marketing pitch here it's just

do the numbers.

I know what you'd have to spend to get 200,000 IOPS and 100 terabytes on NPS, and now you can have your high-speed storage yet still enjoy all the cloud compute capabilities of whatever cloud provider you want.