So another question that I hear customers talk a lot about is this whole contrast of W A optimization versus SDW A. Maybe a comparing contrast is a better way to say it W A optimization has been around for a long time. Um You know, for wide area networks, it's a pretty common application. And so there's some familiarity with that. And so when this SDW a topic comes up many times, it's a, well, I've already got W optimization.
I'm, I'm already doing Sdwan, not necessarily the case though. Uh you know, everything's evolved, right? And there was a time and place when way IZATION really saved our bacon. Um And in order to try to, you know, help illustrate how it saved our bacon. You know, if you just go here really quickly and I'll draw a branch here real quick. I'll draw a data center. All right. Now in your data center, what were some of the things that we were using back in the day exchange Microsoft exchange? Right. Absolutely. We put our exchange there, right? We would have our CRM. Yeah, maybe a big uh er P solution, big er P solution. And what have you. Yep.
OK. Now, this may seem strange for a lot of the new guys out there. But uh old guys like me, these actually used to be in the branch. They were built on a land, they were built to be run in the branch. And so back in the day, you would see all those services here in the branch at some point MP LS came along very narrow bandwidth, but it allowed them to centralize all these applications, which is great. It was awesome, very quickly. We learned those applications were never built to be run over the wan. And some of the problems that we have with the wan right is number one is bandwidth, low bandwidth.
These applications are chewing up a ton of bandwidth. So I have a problem with bandwidth. The other one I have is, you know, we're using TCP, right. TCP was built way back in the day of modems when you know transmissions, right? Were you know, uh people trying to dial up on modems using regular internet connections. So it was really really conservative, right? One of the other things that we were trying to overcome, right, was data redundancy or D data redundancy elimination, right? Trying to pull things off of the net that didn't need to go over these small links.
So and one of the other things too that they also did was application optimization as well. So what we're really trying to do here and everything was predicated on the fact that these links right here are maybe 1.5 megabits, which is crazy when you think about modern networks today, right? So we add SD Wan into the mix and we add big links and this is not a stretch to think that even your most remote branches would have a situation like this. So if I have big fat links, right, with tons of data available to me. Well, then I really don't need W optimization for bandwidth. It's no longer an issue. I've got tons of bandwidth, modern day applications that were written for the web think sales force think office 365 all these applications have been built rebuilt to run on the W A.
So I really don't need TCP W A optimization as well because they've been rewritten. Well, data redundancy of elimination. Most of these applications are now secured. They're all running on a on a secure socket. So if they're running on a secure socket, I can't rebuild right? The application. So it kills my wan optimization. And I can't do dated red density elimination because I can't see inside the packets. So all of these need all these things, right? That would predicate me needing W optimization are all out the door. Now, there's one situation where it still makes sense and that situation is if this is in London, right? And that is in Sydney is in Sydney, right? One of the things we have not been able to figure out how to do is to try to make light go faster.
So if you've got um places that are far across the globe, your latency is still going to affect a lot of these things in here. And so that's what we call mid mile. So way in optimization still has a place in mid mile. But out at the branch, I don't see it having any long term, you know, survivability. So kind to bring this one to a close certain use cases, you know, large geographical distances between it.
There's still probably some value there. But thinking more and more about kind of how much is going into multi cloud and running as SASS and so forth in terms of where that data sits, it's a lot closer to that branch office than you probably think, which means all those other items really are negated. So absolutely great conclusion for this. I think it's been a great example of uh you know, answering one of these key questions on the customer's SD Wan journey SD Wan versus Wan optimization. Hope you found it helpful.
