Email a colleague    

March 2018

Masergy: The Magic of Home-Grown SDN, Hybrid Networks, and Build-It-Your-Way Flexibility

Masergy: The Magic of Home-Grown SDN, Hybrid Networks, and Build-It-Your-Way Flexibility

How do networking breakthroughs come into this world?

Well, in the days of circuit-based telephony, the network equipment providers (NEPs) were the ones who came up with the fresh ideas.  In those days the lines of responsibility were quite clear: NEPs invented network products and carriers bought them.

And, interestingly, this industry model mirrors the way the airline business operates today: the mission of the Boeings is wholly distinct from that of the Deltas, Uniteds and Southwests.

However, in today’s IP-driven telecom world, software has leaped to new heights versus the equipment side — meaning there no longer needs to be a sharp dividing line between operators and solution engineers.

In fact, a few small- to mid-sized network provider shops, with names like Masergy, PacketFabric, Epsilon, and GTT are magicians of network innovation.  Because these firms are not encumbered by legacy networks, they are free to develop and nurture their own software-defined network controllers.

Now even though software is in their genes, the last thing these savvy solution shops want to do is develop and sell off-the-shelf WAN optimization products on the open market.  That would be foolish on their part.  Why?  Because the more profitable route is to keep their software proprietary and embed it in a managed network service they can offer to enterprises (retail) and/or carriers (wholesale).

Plano, Texas-based Masergy is one of the Merlins of the communications business because they deliver home-grown and software-defined networking with cutting edge capabilities.

Now Masergy is not a household name, so I think many will be surprised to know the company did about $350 million in 2017 business.  Impressive.  And rather than go after Fortune 500 ($5 billion-a-year+) firms, Masergy specializes in serving small- to mid-sized enterprises in a variety of industries with its largest customer doing less than $2 billion a year.

In the interview that follows, Paul Ruelas, Masergy’s Director of Product Management, explains Masergy’s unique value proposition: everything from its flexible hybrid networks and customer-controlled portal... to its technology-neutral, “build-it-your-way” consulting and innovative cloud firewall.

Dan Baker, Editor, Top Operator: Paul, it’s obvious that Masergy is a firm that chose the less traveled path.  What’s the company’s business philosophy?

Paul Ruelas: Dan I’d spent years at a US based carrier, developed solutions for cell site backhaul, and worked both in the wholesale/enterprise networks.

And what struck me about Masergy — and convinced me to join them — was that independence and free-thinking.  Our CTO, Tim Naramore, is an interesting guy and before I joined Masergy I heard his major points in a video:

  • “We don’t just launch a product because it’s a check box on an analyst’s column.”
  • “We do not design based on a specific piece of intellectual property from a vendor.”
  • “We develop our designs based on the customer’s requirements and their business challenges.”

Now when Masergy was formed back in 2000, the term “software-defined network” wasn’t quite known yet, but even then we were known for our ability to provision fast and get things done.  We wrote our own network controller system from the ground up, and have modified it over the years.

Today everything we do is done through our console.  No one is logging into individual devices to routers, VLANs, etc.  Our centrally managed controller provisions all the pieces and parts of our network — and the customers themselves can provision things through our portal too.

And one of the things that surprised me when I came here: Masergy was the first non-facilities based carrier/provider I worked for: this means that we don’t have a local plant, copper, or fiber in the ground.  And we are much smaller, of course than Verizon with its 155,000 employees.  We have about 600 people.

Having no access facilities is certainly unusual.  How do you make that work?

Well, we wholesale from over 200 access providers.  And that gives us great control over quality.  If we’ve got a local loop that’s giving us trouble, we’ll swap it out and order another loop.

What’s neat is we don’t buy QoS services from the wholesaler because we have no use for their MPLS.  We want a pure circuit with no QoS, 100% real-time from the customer site to our POP so that we can put our own QoS over the top and customers can control — by themselves — how much voice, video, and data each QoS plane gets.

Bottom line: when customers buy a circuit from us, they don’t get it broken up in pre-defined ways (10% real-time, 20% business class, etc.), and they appreciate that flexibility.

Besides your exclusive use of wholesale access, how else do you differ from your rivals in the enterprise-serving business?

Dan, our customers get a lot of value for the money.  For instance, when they order their primary layer-3 service, they also get an additional two layer-2 services for the same price.  There’s no upcharge to add these — they come included.

Now folks ask, “Well, why would I want that?” So what can you do with these layer 2 services?  Well you can create a VPLS for, say, data replication on layer 2 or a VPWS ELAN service, and even provide public Internet on the same link.  So if your backup Internet should fail, you can fail over to the private side and still get public Internet access.

Our service comes bundled as VLANs.  You stack them, with our Masergy Intelligent Bridge, or you can break them out to physical ports as well.

The big advantage here is the customer can mix and match.

And we also leverage a public loop.  We go and source a private connection to get them back to our network, but if they want to leverage, say, a dedicated Internet access or a broadband connection, they can use that too and connect to the same port on the Internet and now they’ve linked those traditional MPLS services to a publicly served site as well.

We’re also global with tight SLAs that give a flat, pure IP network across the globe where our jitter and packet loss is practically nothing.  Our SLAs also guarantee 100% in-sequence packet delivery for real time apps across the globe.  So if we miss one packet, boom: we need to pay out.  Even still, our jitter from all our POPs — even though it’s global — is less than one millisecond across the whole globe.

We have a 50 millisecond response time of failover, so if they’re on the phone, they don’t even hear the click if the IP signal drops.  It just keeps moving on to the next conversation.

So how is your service different from what a large carrier would offer?

One of the biggest differences is service charges.  A large carrier will charge you for adding a layer-3 service.  Then, if you want a layer 2 or anything else with that, that requires a different access loop, so you need to pay another local line charge.

Our philosophy is: “Hey, it’s a pipe.  And you can do with that pipe whatever you want.”

You don’t have to get another access loop for every single circuit.  So that’s what’s nice about our software: the customer is in control.

Say a customer wants to change a static router or add an EGP (Exterior Gateway Protocol) neighbor to the private network side.  Well, they can log into the portal to make that change.  In a couple clicks they reach the screen where they change their IP address.  Likewise, they can quickly change a network or telco-owned device on the fly.

By the way, the portal the customer uses is the exact same SDN controller our NOC uses.  Our customer sees what our NOC folks see.

Now just because we use a common console doesn’t mean we are cookie cutter.  For instance, our applications at each site are quite different depending on the need.  So if the enterprise CEO and staff is on a particular site, we know they typically need to have both voice and video communications aboard.

Meanwhile, if the site is a kiosk in a mall or we’re serving a remote worker, the application there needs to be flexible enough to scale up and down.  If you try to lock a customer in at a lower capability, that strategy can backfire on you.

What about when the customer needs extra bandwidth?

It’s the same method.  For instance, many customers say, “Look, most of the time I need 20 to 50 Megs but from time to time I may need 100 megs at this remote site.”

So what we’ll do is over-provision a 100 Meg local loop for them that’s clear all the way to our POP, then they pay the contracted rate, say, 20 Meg till they need more.  Then, when they want an extra bandwidth burst, they have a couple choices:

  • Instantaneous dial-up of the port speed to anything between 20 and 100 Meg — with a notice about what the added price per day is going to be for that circuit; or
  • Schedule the bandwidth for a particular time slot.  Say they’ve got a backup they want to schedule the third Thursday of every month.  So they can specify the exact hours of that backup and have it automatically step up and step down the bandwidth — once again, informing the user of its costs.
So what is your sweet spot as far as customers are concerned?

We have a couple of different ones.  We attract the value buyer because we don’t nickel-and-dime customers with service charges.  Basically any change you make, there’s no service charge.

Then we also appeal to the tech-savvy folks who want to run their network themselves and be on the cutting edge.

And of course, when you serve the tech-savvy guys, you have no choice but to constantly evolve your technology.  You need to be flexible, agile, adapt and optimize.  The worst thing would be to leave our customers behind the capabilities curve.  But it’s tricky because we can’t nickel-and-dime them either, so we need to be efficient.

An example of adding new capabilities was our December announcement about tying SD-WAN products into our SDN controller.

What about cloud connectivity.  You have providers out there like Megaport who specialize in that.

It’s interesting.  We have some similar services to Megaport, our Cloud Marketplace where customers get a dedicated connection right into the SaaS provider.  We are inside almost every Equinix POP on the globe and we interconnect with them for AWS, Azure, all those providers, but we also tie it back into the network for SLAs.

Megaport provides a good service — nothing wrong with it.  Where we differ is in offering a broad number of services that a customer may be looking for.  That’s important because you want to tie the network piece into cloud connectivity.

I run the network side where we have three basic product sets:

  • Hybrid Network capability, as we’ve discussed: cloud connectivity, private networking, 4G, and Internet;
  • Unified Communications with cloud contact centers, SIP trunks, and Cisco IP phones; and,
  • Security Services such as managed detection and response, managed firewalls, and managed routers.

So for customers looking for one-stop shopping, we bring it all together as a solution, one set of trouble tickets, and one portal.

Tell me more about your security services.  What’s noteworthy there?

In the security area, the enterprise needs to be careful especially as they add new network technologies or merge with another company and bring the new network inside their firewall.

What we do is allow customers to spin up upwords of 4,000 VLANs — more than they will probably ever need — and protect everything with our Cloud Firewall so they can test those new networks safely.

For example, a customer came in today and said, “Look, I’m a current customer of yours, but I just bought a company in the U.S. that has eight sites and they only have internet.  So how can I bring them into our private network?”

Well, I said, “Let’s go to the portal and you can order up Network as a Service or SD-WAN right there, and you can turn up your own devices on all those sites and put them in their own layer 3 cloud.”

So the Cloud Firewall sits between their network and the new acquisition network.  This enables you to bring them “on-net” so to speak: see all the sites and connect.  But the neat thing is the Cloud Firewall keeps them at a safe arm’s distance.  This is key during the transition period when you need to monitor and make sure the right security policies are there and the new network is not infringing on the home network.

There’s a great debate in the market over MPLS versus SD-WAN technologies.  What’s your perspective on the issue?

Dan, I frequently get this question.  And I tell people: it’s not a binary choice but a mixture.  The term hybrid networking makes sense: choose the appropriate circuit for your need.

Folks choose which side of the risk tolerance line they want to be on.  On one side you’ve got private connectivity, such as MPLS, and that will cost more but offer a very low risk of failure.

At the other end of the spectrum you’ve got broadband, cable and DSL that are more risky but come with a low price.

Then, in my view, about two-thirds of the way to MPLS is DIA, Dedicated Internet Access, which is still public access, but it’s really private access on a public loop.

I tell people, “Draw the line and tell me where your sites fit in.”

Today, most of our enterprises are putting in MPLS as their primary, either 10 or 20 Megs.

And they put in broadband or DIA to back it up and then use SD-WAN over all of it, so they achieve a kind of best of breed.  They can put their voice and video on the highest quality primary line then fail over to the secondary should they need to.

Now Masergy has had WAN optimization and performance analytics for a long time, so we’ve had the piece parts that make up an SD-WAN.  To me, most current SD-WAN deployments are a peripheral product set in an enterprise.  And as the months and years go by, you’ve steadily seen this capability folded back into the network.  So the networks will give you the same attributes of the SD-WAN — only it will not be dependent upon a particular technology partner as it is today.

For instance, our Intelligent Network Analyst gives customers the visibility to see when a site goes down and view on-going performance whether it’s at the application layer or the IP layer, the ports, or QoS.  They can also slice and dice the data on the live network.

So you’re being technology neutral and queuing up flexibility as the enterprise moves forward.

That is what we like because we don’t know what the customer wants to do.  Heck I’m surprised what they do sometimes.  We say, “Whatever you want, this is your network, across the globe.  If you need more bandwidth, a different type of bandwidth, or some application-specific stuff, let’s do it.”  That’s how we work.

We don’t lock the customer in with our own pre-defined deck of services, speeds and feeds.  We sit down at the white board and say, “Tell us about your sites, tell us about your applications, and let’s figure out what sort of network you need.”

Last year, one of our largest customers came to us and said, “Give us an SD-WAN presentation.” So, we came up there and said, “Look, we’ll show you our view of SD-WAN, but let’s also work the whiteboard.” Sure enough, at the end of the day they decided to get pure MPLS and a NAS Internet backup without SD-WAN for now.

So sometimes we talk customers out of new technology, not because it is bad.  We say, “Look, I will sell it to you, but if you want us to solve your problem, this particular service is probably better for you.”

Paul, many thanks for this fine briefing.  Looks like Masergy is on a nice roll.  Look forward to keeping up with your progress.

Copyright 2018 Top Operator Journal

 

About the Experts

Paul Ruelas

Paul Ruelas

Paul Ruelas is Director of Network Products at Masergy.  He brings over 26 years of expertise in telecommunications, IP Networks, complex solution design, and product development.

Paul has developed many Ethernet and optical products that enabled numerous global enterprises to transform their data communication infrastructures to improve business outcomes.  Paul is an industry thought leader in communication transformation and on topics such as hybrid networking, SDN, NFV, and cloud connectivity.   Contact Paul via

Recent Articles