Governing the Wild, Wild WAN
The latest BriefingsDirect podcast discussion focuses on the rapidly escalating complexity and consequent need for network management innovation in the age of hybrid computing.
And that's because the emphasis nowadays is on "networks," not "network." Long gone are the days when a common and controlled local area network (LAN) served as the workhorse for critical applications and data delivery.
With the increased interest in cloud, Software as a Service (SaaS), and mobile computing, applications are jockeying across multiple networks, both in terms of how services are assembled, as well in how users in different environments access and respond to these critical applications.
Indeed, cloud computing forces a collapse in the gaps between the former silos of private, public, and personal networking domains. Since the network management and governance tasks have changed and continue to evolve rapidly, so too must the ways in which solutions and technologies address the tangled networks environment we all now live and work in.
Automated network unification and pervasive wide area networking (WAN) governance are proving essential to ensure quality, scale, and manage security across all forms of today's applications use. Join us as we explore the new and future path to WAN governance and to better understand how Ipanema Technologies is working to help its customers make clear headway, so that the next few years bring about a hybrid cloud computing opportunity and not a hastening downward complexity spiral.
We're here now to discuss the new reality of networks and applications delivery performance with Peter Schmidt, chief technology officer, North America, for Ipanema Technologies; and David White, vice president of global business development at Ipanema. The panel is moderated by BriefingsDirect's Dana Gardner, Principal Analyst at Interarbor Solutions.
Listen to the podcast (33:49 minutes).
Here are some excerpts:
Peter Schmidt: As soon as you start using multiple networks, you're in the cloud, because now you're making use of resources that are outside the control of your own IT organization and your service provider. Whether people think about it or not, just by adding a second network, they're taking their first steps into the cloud.
Anybody who carries a smartphone is experiencing the personal, private, public boundary of operations themselves. But what seems natural to somebody carrying an iPhone or Blackberry is a tremendous challenge to the traditional models of IT.
Even as little as three years ago, the focus was on how to get the most performance for your applications out of your single MPLS network. I am talking enterprises where all of their applications are hosted on their property. They've got a single MPLS network from one service provider and they're still struggling to deliver reliable application performance across the infrastructure.
Now, we throw in multiple places to host applications. You have SaaS, Salesforce and Google Docs. You have platform as a service (PaaS) and infrastructure as a service (IaaS). People's critical applications can be hosted in numerous locations, many of which are beyond their control. Then, as I mentioned, these are being accessed via multiple networks, and you have the legacy MPLS plus the Internet.
There are increasing numbers or diversity of models of those networks, whether the Internet connection gets to a service provider POP and then via MPLS to their own data center, or what is the impact of content delivery networks? So we've got a situation where enterprises who are struggling to master the complexity with one data center and one network are now using multiple data centers and multiple networks. Something is going to have to give.
David White: This is also all focused once again on the branch office. We've had server consolidation where we try to remove any type of issues for the branch and remove intelligence from the branch. As cloud computing has come in, we are now putting more stress on the branch.
We're not necessarily putting intelligence out there, but we're having two, three, four, five or more networks, all coming into the branch at the same time, and that traffic has to be managed. It's something a lot of people haven't thought about.
When you look at the announcements that have been coming out and the hype on cloud in the industry, it's all focused on the data center. That's because most of the vendors say, "That's where the big bucks are being made. We are going to make money out of the data center."
Ipanema, on the other hand, is focused on application acceleration, and in order to do that, you have to take care of what goes on in the branch and manage it.
At a high level, the first thing you have do is provide some type of WAN governance, simply meaning that we are going to make sure that you have taken care of the management of your business. Because that's what WAN governance means -- providing the type of control over your business to allow it to continue to be productive, as you're making changes to your WAN.
Simply put, you first of all have to find out what's going on in the network. You have to understand what's happening on those four, five or six different flows that are all going in from different sources to your branch. You have to be able to control those flows and manage them, so that you don't have your edge device or edge router getting congested.
You have to be able to guarantee performance and, very importantly, you also have to then unify, balance, and optimize the performance over those multiple network points that are coming into your branch.
If you're doing it the right way, at least what we would say is the right way, it needs to be dynamic, automatic and, in Ipanema terminology, autonomic, meaning that not only does it happen automatically, but the network learns and manages itself. It doesn't require extra human intervention.
Schmidt: The way the enterprise is going to get its arms around this increasingly complex environment is not through throwing people at it. Throwing people at network management has never worked and, now that the environment is more complex, it's going to work even less.
The whole point of cloud is that you're going to use virtualization and automation to bring up instances of servers quickly and automatically, and that's where this order of magnitude improvement potential comes from. But if you don't want the multiple networks to be the bottleneck, then you have to apply automation in that domain as well. That's what we've done. We've made it possible for the network to run itself to meet the businesses' objectives.
The effect that has in a branch office with multiple network connections is really to hide all the complexity that that multiplicity brings, because the system is managing them all in a unified way. That's what we're getting at when we're talking about network unification. The details that bedeviled traditional management just kind of disappear.
WAN governance is what the CIO wants to buy. CIOs don't want to buy a WAN, and they certainly don't want to buy WAN optimization controllers. What they want to buy is reliable application performance across their infrastructure with the best possible performance and lowest possible cost. My high-level definition of WAN governance is that it's the technology and techniques that allow the CIO to buy that.
We're about to release our first iPhone app to provide an interface into our central management system, and it's terrific. It's exactly the kind of thing the CIO would want to have in their hand. That just shows the value of pushing IT to be democratized and put into the hands of all of the people tied to the enterprise.
The Ipanema system is designed to provide the full control by giving the enterprise IT organization, not just visibility in reporting on every user's access to their IT infrastructure, but also to automatically control all of that traffic in accordance with various policies.
We don't see any other way around it. You're not going to do this manually. You've got to build smarter systems. We happen to think that we are a huge piece of that puzzle in terms of how we control things at the network level.
White: And we look at this WAN governance as really a piece of ISO standard for IT governance, which is an official ISO standard. There is a section in there on WAN governance. In a way, it talks about what you have to do to manage your wide area.
Ipanema strongly believes the WAN governance is really a standard that should be put on the books, but isn't yet. If you're really going to have governance over your IT, since the network is a strategic asset to promote enterprise customers, you need to have governance over the wide area as well.
Schmidt: Ipanema has pioneered a unique approach that stems from the idea that all that matters is that end users are able to get good performance from their applications, because that's when they are most productive. When application performance slows down, end users start surfing the Web. So ensuring the performance of the application is critical. That's what the enterprise needs to reorient itself toward.
The fundamental input into our system is a list of applications and their performance. The system itself is intelligent enough to monitor and dynamically control all of the traffic to achieve those objectives on behalf of the business. So it's imposing the business's will on the network.
White: It starts with our three founders who got together and took a look at what the needs were from an application perspective. Their goal was to find a way to ensure that, as users, we all had the performance we needed and that enterprises could deliver performance from an application perspective to their users.
That's what they started out with. Then they took a look at how you would deliver that service and recognized the best way to provide for the delivery of the right type of consistent application performance is to do it over the wide area and to look what happens over the WAN.
They were very visionary in recognizing that application performance over the wide area is going to be the single most critical piece of the puzzle, when it comes to taking a look at how we as users of enterprise deliver service and do it in conjunction with major service providers and network providers, because they are the ones that deliver the wide area connections.
When they started out, they were told that they were wrong and weren't looking at it the right way. When you see what's happened to the network and how it's evolved, particularly now that we are moving into the cloud generation, they were focused exactly in the right area. Although we have a lot of new features, the basic architecture has been there for years and it's been proven in major service provider networks and is installed on a global basis.
Schmidt: There are a couple of things that are the secret sauce, but the easiest one to explain probably is the fact that our appliances actually cooperate with each other, and this is unique. Our appliances know about not just the traffic that's impinging on their network interfaces, but they actually know about the flows that are active everywhere on the network.
It's actually not that that simple. They really only need to know about the flows that might conflict with the flows that they are managing. But conceptually, every device on the network knows about all the other flows it needs to know about. They are constantly communicating with each other -- what flows are active and what performance those flows are getting from the infrastructure, which includes the whole WAN, but also the data center and the service. So what does that enable?
Sharing this information means that all of the decisions made by an individual device are made from a global perspective. They're no longer making a local optimization decision. They each run the same algorithm and can come to the same result. And that result is a globally optimum traffic mix on the network.
When I say globally optimum, that's a valid technical term as opposed to a marketing term, because the information has been collected globally from the entire system. In terms of optimum, what I mean is the best possible performance from the most applications using the given network infrastructure and its status at that point in time.
White: The point I'd like to make is that it's absolutely impossible to measure it in a cloud environment as an enterprise network manager, because you only see a piece of the network. Unless you've done something different, which is what we provide, than the way you are going to look at your network, if you are looking at it the way you've done for the last 10 or 20 years, there is no way that you can see everything.
The closing point here is that the first step is visibility into the network, and the next step is providing the control. You need to do that in the cloud environment, and that's what Ipanema does.
Dana Gardner is president and principal analyst at Interarbor Solutions, which tracks trends, delivers forecasts and interprets the competitive landscape of enterprise applications and software infrastructure markets for clients. He also produces BriefingsDirect sponsored podcasts. Follow Dana Gardner on Twitter. Disclosure: Ipanema Technologies sponsored this podcast.