How ‘hyperscalers’ – such as Google and Microsoft –are using Open Networking to scale in the enterprise

(Image credit: Shutterstock)

Networks are growing, and growing fast. As enterprises adopt IoT, VPN technologies, virtual machines (VMs), and massively distributed compute and storage, the number of devices—as well as the amount of data being transported over their networks—is rising at an explosive rate. It’s becoming apparent that traditional, manual ways of provisioning simply don’t scale.

Something new needs to be used, and for that, we look to hyperscalers; companies like Google, Amazon and Microsoft who’ve been dealing with huge networks since almost the very beginning.

The typical approach to IT operations has been focused on one server or VM at a time. Any attempt at management at scale frequently comes with being locked into a single vendor’s infrastructure and technologies. Unfortunately, today’s enterprises are finding that even the expensive, proprietary management solutions provided by the vendors who have long supported traditional IT practices simply cannot scale, especially when you consider the rapid growth of containerisation and VMs that enterprises are now dealing with.

In this article, I’ll take a look at what organisations can learn from this approach, and how they can use open, scalable network technologies—those first created or adopted by the aforementioned hyperscalers—to reduce growing pains.

From hyperscale to enterprise

Many organisations have a traditional network architecture designed mostly to handle north-south traffic, with very little attention paid to east-west communication between workloads. But that architecture is becoming less and less popular as microservices are adopted, and applications become distributed. However, this presents a bit of a problem—namely, how does an organisation scale out their network when traffic patterns within data centres are changing in ways that challenge the very design principles of today’s networks?

The answer is open networking. Open networking consists of standards, protocols and technologies that are open, extensible and not tied to a specific vendor. These technologies allow your network to scale dynamically, adding hardware, software and services to meet the specific needs of your organisation. Open networking uses an open source-based network operating system that can be installed on a wide range of different switches, allowing organisations to choose the hardware that best suits their needs.

To see an excellent example of open networking and open scaling in action, take a look at the open networking giants: hyperscalers. These companies have had to deal with rapidly changing networks with brand new technologies being added every day. The networks of the hyperscalers are highly automated and make extensive use of APIs. Their infrastructure and applications are deeply integrated, enabling not only the rapid and efficient scaling of their network infrastructure, but of the applications supported by that infrastructure.

Companies like Google, Amazon and Microsoft are excellent examples of how an organisation can use open source software and networking technologies to meet a diverse range of scalability and automation needs. Fortunately, the approaches they’ve used to achieve success also scale down, as well as up, and can be applied in an enterprise environment.

Scale requires automation

Open networking has trickled down into almost every pricing segment, allowing the smallest start-up to afford an automated, scalable network that would’ve been unrealistic to consider even 10 years ago. The time is right for organisations of all sizes to start looking into when and how they can move to open networking; almost every business sector in the world is going to face a problem scaling their networks, and it’s better to get ahead of the problem now than wait for it to become intractable.

One of the key benefits of open networking is that standardised APIs make automating, scaling and integrating your network architecture with your applications easy, using any programming language or automation tool you care to name. Network automation at scale also enables all your devices, tools and programs to talk using open and standard interfaces.

Many organisations—both large and small—are adopting open networking and web-scale architecture. Together, these technologies can, pursuant to the organisation’s requirements, create a self-healing, software-defined, fully open network, which adapts to changing conditions. They can also reduce total time at the console for system admins and increase productivity across the board.

A modern management platform is also a benefit here. Platforms like Cumulus NetQ for example build on open networking’s strengths to provide a next-generation management solution that brings hyperscale management to the enterprise. Traditional technologies, such as SNMP and the router/switch console of course still have their place. However, if an organisation wants to scale to the next level of networking they need to investigate migrating to open, scalable and secure networking as soon as possible.

Adopting open networking technologies at web scale has been shown to reduce an organisation’s CapEx by one-third, when compared to single-vendor solutions. It’s also been shown to reduce OpEx by 75 per cent by optimising the organisation’s operator-to-switch ratio. Additionally, a time-to-production drop of 95 per cent was observed in most cases, leading to massively increased operational efficiency. As organisations and networks grow, the impact of those percentages and ratios become more meaningful, resulting in both significant money and time being saved.

Pete Lumbis

Pete Lumbis is one of the first 100 CCDEs (Cisco Certified Design Expert) globally and the first in the global Cisco TAC. Focused on customer success and technical excellence, Pete is a go-to expert on MPLS, multicast, IPv6, EIGRP, OSPF and BGP. Pete has been a presenter at Cisco Live and a guest lecturer at North Carolina State University.