Duncan Moore

You called attention to it.

So even in a massively distributed environment, so let's say you take a namespace and spread it across 16 data centers and you've got 100 billion things in there and I do a put, I am read after insert or read after write 100% consistent across the entire namespace.

So, you know, you can do different consistency levels with storage grid as well.

So, you know, if I do an overwrite of an object, that's not going to be read after write across the entire namespace, right?

Because I could potentially have older metadata cache somewhere.

But for a new object, it's 100% read after write.

But, yeah, the…

The short answer is that that lack of an update kind of helps things out a lot.

Yeah, so I mentioned at the beginning I'm old, right?

I've been here a while.

So when we launched SnapMirror in –

1999, with data on tap 5.3, codename Bex, for those that remember and have the shirt, right?

I remember running around, and I was the NDMP guy, too, going, tape is dead, right?

Still here.

It's still here, right?

Files are not dead, right?

I mean, there's, you know, there's a lot of use cases that object isn't the answer.

But the ones where it is, it's the only one.

I mean, it's solving problems that you just can't solve any other way.

I'll go head to head with the big S3 any day of the week.