Building a Network for the Brontobyte Era

If your company is running on a legacy network, this story will be all too familiar. From mid-June to mid-July of this year, hundreds of millions of people worldwide tuned in for the World Cup. Here in the U.S., many of the tournament’s 64 matches fell right in the middle of the American workday.

You would have seen it if you walked into any office that month—lots of employees wearing headphones, furtively watching live video streams of faraway soccer matches on tablets, smartphones, auxiliary monitors and in nested browsing tabs.

What most people didn’t know was that legacy networks across the country were groaning under the weight of that unexpected traffic.

I heard stories of executives trying to connect to HD video conferencing calls, and not being able to get onto their own network. Critical business functions—financial transactions, multimedia file transfers, customer service contact centers, and more—were either impacted, or crashed entirely.

Those failures weren’t in the applications. They were in the network.

Legacy networks were designed for a different era, with myriad boxes and switches that have to be manually configured and networked with one another.

Traffic hits bottlenecks in the system, leading to slower speeds, and in some cases, total outages. Getting the network back online is a physical job, run by technicians inside the data center. Upgrading equipment requires planned downtime.

Adding new applications, such as HD video conferencing, can take weeks of step-by-step configuration and careful testing. After all that work, 80 percent of companies deploying a new application will fail on their first try, due to complex network configurations leading to unexpected problems.

Networking demands aren’t getting any easier—it’s estimated that global IP traffic will reach 131.6 exabytes per month by 2018. It’s difficult to wrap your brain around how much data that really is, but by one estimate, 5 exabytes is the equivalent of every word ever spoken by mankind [].

By 2018, the world’s networks will handle 5 exabytes of data every 28 hours. It won’t be long until we’re measuring network traffic in zettabytes, yottabytes and brontobytes.

Here at Avaya, we’re building future-ready, virtual networking software and appliances for the brontobyte era.

Avaya Fabric Connect is a virtualized, software-defined network designed to be flexible, scalable, and easy to set up and manage.

Using Fabric Connect, companies can give specific applications priority status on the network, making sure, for example, that critical business functions like HD video conferencing and financial transactions get a green light, while YouTube and Facebook take a back seat.

A related product called Avaya Fabric Attach makes it easy to automatically add new networking endpoints—everything from a new router or switch to an IP-enabled security camera. As new endpoints get connected to the network, Fabric Attach automatically identifies and provisions services for those endpoints.

Now, the network is a single entity, rather than a series of individually-programmed boxes.

Unlike many of our competitors, our network architecture was built from the ground up on open, IEEE standards. Avaya helped author those standards, and our engineers continue to contribute their intellectual capital to both the OpenStack consortium and the OpenDaylight project.

Fabric Connect and Fabric Attach got their biggest tests to date earlier this year, when Avaya built the network in Sochi, Russia that powered the 2014 Winter Olympics. Every live video broadcast out of Sochi flowed across Avaya’s network, as did every gigabyte of WiFi data from the 120,000+ mobile devices brought to the Games by athletes, fans, journalists and staff. The network ran with 99.999% uptime.

Network virtualization offers a path forward for companies struggling to keep up with growing data demands. Open, flexible, standards-based technology means the network will be able to handle the types of bandwidth-hungry devices and applications of the future that haven’t been invented yet.

Related Articles:

Avaya Pulls in Two Nominations for "Best of Interop" Award

“It’s an honor just being nominated.”

Academy Awards TrophiesFeatureflash/Shutterstock

How many times leading up to a prominent awards ceremony–whether it be the Oscars or the Grammys–do you hear that statement?

As we enter this year’s Interop Las Vegas event, I am going to be bolder and say that we are honored to be nominated–not once, but twice–for the Best of Interop award. But we really want to win!

Avaya was shortlisted in the Networking category with a new feature called Fabric Attach, and in the Security category, with our Identity Engines product.

In this blog post, I am going to focus on Fabric Attach, a feature many of you may not be familiar with yet. At the highest level, Fabric Attach extends the benefits of the Avaya Fabric Connect architecture (based on an enhanced implementation of IEEE 802.1aq Shortest Path Bridging) to non-fabric-enabled devices and end points. It automates the identification and provisioning of edge devices or end points (phones, APs, cameras) that connect into the fabric.

So what’s the value? It really comes down to faster onboarding of users and devices with less chance of error. In a recent Avaya press release that looked at network downtime, it was found that 82% of companies experience network downtime as a result of errors when making network changes.

More dramatically, change-induced network downtime can have serious career-limiting consequences, with 20% of businesses acknowledging dismissals. By combining Fabric Attach’s “zero touch” edge provisioning, and device onboarding with the ability to “set and forget” your Fabric Connect core, the risk of downtime is dramatically reduced.

Even more importantly, the network engineer’s job suddenly becomes less burdensome with less manual provisioning and network firefighting. They can now devote their time to activities more strategic in nature.

Avaya Fabric Attach

To attach non-Fabric Connect-enabled Ethernet switches to a Fabric Connect network, it’s as easy as taking the switch out of the box and plugging it into a Fabric Connect-enabled edge switch, so that it can automatically configure itself with fabric-based services that have already been pre-configured in the fabric core.

Fabric Attach also allows for “zero touch” onboarding of end points. This is where our lessons learned from Sochi come into play. If you have been following Avaya and our role in the Sochi 2014 Winter Games, you’ll remember that it was one of the largest BYOD/guest environments in the world.

One of the challenges we faced was providing services for 40,000 different users, all connecting simultaneously, the day of the Opening Ceremonies. Identity Engines played a critical role in authenticating the devices and then dynamically assigning them to the correct VLAN based on the credentials.

That VLAN was then assigned to the correct virtualized network. What we are doing is taking this concept one step further and dynamically attaching the end point directly into the virtualized network. It’s basically adding a policy to Identity Engines that allows the network to create the VLAN (or SSID if its wireless), create the virtualized network (or I-SID if you are familiar with SPB) and map the VLAN/SSID to the virtualized network (or I-SID) all without user intervention.

This capability would have even further simplified things for the Sochi team, especially considering that we had to plan for 120,000 wireless devices connecting at peak periods such as the Opening and Closing ceremonies.

Avaya Fabric Connect

So although the Interop judges have some tough decisions to make as to who they select as their winner, we are honored to be nominated, and we still want to win!

In my next post, we’ll talk about Identity Engines and why this product is getting attention from the Interop judges.