(lively electronic music) >> Announcer: It's The Cube! Covering Sapphire Now 2017. Brought to you by SAP Cloud Platform and HANA Enterprise Cloud. >> Okay, welcome back, everyone, to our special Sapphire Now 2017 coverage from our studio, I want to thank SAP Cloud Platform and SAP Enterprise Cloud for sponsoring our three day coverage, to allow us to go and do some original content. Also, we have Google IO going on, some coverage there, we'll get some of those segments on also, other events going on all around tech. Wanted to bring in to this studio session for discussion is Bill Norton, a CUBE alumni, the chief scientist at Console Connect, also known as Dr. Peering, a real expert in networking, and how networks are built and how they run and the impact of how packets move around the network, and of course, networks are being abstracted away to software and of course, no better person to talk about software-defined networking, the changes in the network landscape, and more importantly, the security implications. Bill Norton, here inside our CUBE conversation, it's good to see you again. >> Great to see you, John. >> So, you've got a couple whitepapers, I wanted to call you on twofold. One: a lot of CIF going up the network, okay, as much as people are moving up the stack as fast as they can, the network still has got a lot of action going on. Software-defined, scaling issues, now security issues, still an important part, you've still got to move packet from point A to point B. Now things in the cloud have certainly escalated the conversation. >> Absolutely. >> John: Your thoughts. >> We're at a crossroads right now. We have two forces that are colliding. The first is increased dependence on the cloud, as more companies are using cloud for business-critical activities, as in, you can't take orders if you can't access your Salesforce.com service, or maybe you have cooperating colleagues in different parts of the planet, that are using a shared storage system, right, and they can't cooperate, they can't do their thing unless they can get access to that infrastructure, so there's great dependence on these cloud-based applications, which is clashing with the continued vulnerabilities of the public internet. >> So the internet's at stress, it's like changing the airplane engine out at 35,000 feet. A lot of stuff's going on. What's the bottom line? What are the core issues right now that people are facing? Is it just availability, or is it just bad packet movement, bad network design? What's the core issue right now facing the stress on the internet? >> You know, it's funny, I work with customers all the time, in my daily job, and the number one reason that they come to Console Connect is for security reasons. It's so funny that every single time it's security reasons. They want to be directly connected to those business-critical or mission-critical applications, and the reason is clear: if you're directly connected, the shortest path between two points is a straight line, right, so you get better performance and better reliability. >> Unless the backhoe hits the link and then you're dead, anyway, that's physical connection. So before we get further, I want you to take a minute to explain Console Connect, 'cause I just want to get that out of the way so we can set the context >> Sure >> for the conversation. >> What do you guys do? >> I know what you do, but share with the audience what you do. >> Yeah. Console Connect is a cloud interconnection company. We believe that what the internet needs is to have the ability to directly connect to mission-critical destinations, whether that be a storage service, or an infrastructure-as-a-service company, maybe it's a particular SaaS application. If you depend on it, if it's business-critical (and when I say business-critical, what I mean is, if it's down for a couple of hours and you can't access it, then your business is adversely impacted, so that's what I call business-critical.) If it's business-critical, you ought to be directly connected because you can mitigate against the concern of not being able to access that particular service. >> Alright, so now you have also two things that you've brought as props, whitepapers. >> (laughter) Yeah. >> Let me just see those for a second. >> What's the first one? >> This one is called 'Cloud Connections' and this is a paper I did when I started initially doing some research on how cloud systems work. This is a comparison between the interconnection regimes of AWS, Google Cloud and Azure, and what I found in the research is that these interconnection mechanisms are completely different from one another. It's so interesting that they not only chose different names for all of their services, but in the case of interconnection for direct connections, they also chose a different model for interconnecting. So for example, Amazon Web Services has what's called Direct Connect, and I talk about how that's set up, and both a single and a redundant path to access into your AWS resources. But if you look at Google Cloud's interconnection (they call it Google Cloud Interconnect) and that's really a peering with propagation, it's kind of like peering with somebody, while leaking those routes to the other sides. >> John: Yeah. >> So the Google Cloud Interconnect partner is the one who's propagating that, essentially P2P peering across the infrastructure. And Azure's completely different. Azure decided they wanted to have two bundles of three-tuple peering relationships. One for private resources you can't access over the internet, one for public, which you can access over the internet, and one for Microsoft services. So I laid all this out >> And they call that ExpressRoute connectivity provider? >> That's right. >> I think that's the word you used here. Alright, so let's bottom line, we now have four, well okay, three major horses on the track. Is there any other ones in the air? >> Oh yeah. >> Or shall we just look at the three ones you did? >> I chose those three because AWS, according to Gartner is, whatever, 14 times larger than the next 10 competitors combined, so that's an obvious one to put in there. But Google Cloud and Azure are making strong >> And they have good tech and they've just had their event, too. I did watch that network diagram, Amazon does share a lot. James Hamilton shared a lot of information at re:Invent. Look, he's so popular the phones are ringing. Do you want to grab that? >> No, that's okay (laughter) >> You sure? Is your wife calling? >> Probably, yeah (laughter) >> Alright, so bottom line here is with the horses on the track, three horses, Amazon, Google, Azure, who's better? >> You know, they're just different, they're just different. >> Come on, pick one. >> I think it's really >> Pick one, if you had to bet your life on one which would it be? >> It depends what resources do you need to get access to, it really does. >> Maslow's hierarchy of needs, go. Google, Amazon or Azure. >> I would observe this, I'll get in trouble for saying this but I would observe that AWS seems to appeal to the broad IT crowd. If you're an engineer, you're like me, you like to hack code, you might want to play around with the Google Cloud platform 'cause that's pretty neat. It was designed by engineers so you can see what their mindset was when they designed it. >> Yeah, and YouTube is not a small operation. >> Oh no, no, not at all. >> I mean it moves a lot of traffic. >> Yeah. >> So then, Google knows their traffic game. >> Oh yeah, and of course, massive scale for all these guys. But I also like Azure for .NET types of activities, and of course you can do any workload on any one of these. >> So what you're saying is, some run great on the dry days, some 1k and the horses run good on the rainy days, so different horses, different courses, strengths, >> Bill: Yeah. >> John: weaknesses. As Steve Awodney says, horses for courses. >> You know what the really interesting thing is, is when you can connect those together. We just announced over at the ITW conference our CloudNexus product that we couldn't be more excited about. >> Explain that, what does it do? >> Yeah, it's a mechanism for you to take your availability zones or your regions in one cloud and be able to send traffic through the CloudNexus to get access in to another cloud or another region or another availability zone, and do so without having to go all the way back to your corporate datacenter to exchange that traffic. >> Got it. So basically not a lot of latency. >> Not a lot of latency, it's going to be direct, it's not going to be intermixed with other traffic, it's a internet bypass just like our (mumbles) >> Okay, what's that other paper you have here, let's talk about that one. >> Yeah, just real quickly, this one's called 'The Emerging Private Internet'. >> Okay, how to build your own private business ecosystem for business-critical... and by the way you're the author of both of these, so I'm just going to hold it up there. True private cloud marketshare is going to be in the billions of dollars, Wikibon just put out a study on that. >> Yeah. >> So, hybrid and private clouds are still going to be there heavily with duty, so the notion of private is going to be around. >> Yeah. >> What's this paper find? >> This is a little bit different, this is in response to the denial attack of April 21st or 22nd, I forgot which it is. When that attack happened, that took down dozens of very popular websites that depended on that DNS service. >> Yeah. >> And talking with a lot of the folks that were affected, they recognize that in order to not have that problem happen in the future, they're going to have to have direct connections to those mission-critical destinations, so no matter what happens in the public internet, their critical interconnections and their critical data paths are still solid and robust. So this is the private side of the internet, where you tie down, you essentially lock in connectivity to those mission-critical destinations and the public site is still where you deliver your traffic out to all those who can really get access. >> After a while, yeah, access to all the routes that can be manipulated and or policy-based manipulation as well. >> Yeah, so this is a little bit of a trend now that we're seeing, as companies are migrating over to this new platform. >> I mean, it's like a whole new virtual roadway's got to be reconstructed. I'm a lot fascinated with what... you know, you like to geek out on wide area network stuff, in the past we have. But to me, I think that whole WAN, wide area network concepts from origination to final destination on the packets path is really complicated with cloud, and now we have multi-cloud. >> Bill: Yeah. >> How do you see the routes and the maps and all the naming, all the addressing evolving to support the capacity? >> It's an interesting question. I don't think my crystal ball is any better than yours is. >> John: Well you've got two good papers here. >> I do think we're migrating towards two halves of the internet, the public internet which everyone knows and loves, and gives you great connectivity globally, but that traffic that is attacking somebody else is traversing the same routers and links that your traffic depends on, so I think we're going to see >> I want a secure route that's highly patrolled, and without bad guys. >> And it's solid, and it's robust, the traffic is not intermingled with others, so your traffic is not impacted by the traffic of others. >> Console Connect is your company, had a couple name changes but you guys have grown really fast, you do have a great team by the way, but you're just going, doing your business and it's successful, I've got to ask you what's the main reason why people come to you guys besides your awesome technical knowledge and ability to write great whitepapers and be Dr. Peering, well-known in the industry. Seriously, why are people coming to you right now? Is it because of the alternative? What's the main reason? >> The main reason, as I said before, is security, the fact that any traffic that's being used to DDoS against somebody else is traversing the same routers and links that you depend on. So your traffic, fundamentally in the public internet, is intermingled with other people's traffic, and therefore your traffic can be impacted by other people's traffic. For the mission-critical traffic, you want to have that go across no intermingling elements, so you'd like to have that be directly connected. So security is the number one. Number two and number three kind of go back and forth, it's either better performance or better reliability depending on what issues that particular customer is facing. >> Alright, Bill Norton, chief scientist, Console Connect here inside theCUBE for special coverage of SAP Sapphire 2017. Thanks for watching and stay with us for more after the short break. (lively electronic music)