DataCenterNews Asia Pacific - Specialist news for cloud & data center decision-makers
Story image
Equinix: Localise and optimise traffic at the digital edge
Thu, 27th Apr 2017
FYI, this story is more than a year old

This is the second in a series of blogs on leveraging an Interconnection Oriented Architecture (IOA) strategy for network transformation.

The blogs in this series prescribe the steps necessary to re-architect your network infrastructure by extending the LAN and integrating the WAN at the digital edge for greater performance and a higher user quality of experience (QoE).

The digital edge is where commerce, population centers and digital ecosystems meet. This is the place where information is exchanged, transactions are conducted, and data is gathered and analysed.

However, the inherent limitations of legacy networks and the laws of physics impair high-speed, low-latency communications over long distances. By localising and optimising traffic at the digital edge, you improve network efficiency and performance, reduce risks and lower costs.

The latency challenge

Today's digital businesses are bombarded from all angles and need to manage a myriad of digital transactions, including but not limited to:

  • Harnessing dispersed sources of data for analytics
  • Expanding market reach regionally or globally
  • Connecting mobile customers and employees to IT services and assets
  • Collaborating with partners and integrating with supply chain workflows
  • Accessing existing and emerging cloud, content, IoT and B2B platforms

Latency is increasingly becoming a problem digital businesses can't avoid. You can always ramp up bandwidth over an internet transport, but the laws of physics dictate that you have limited control over latency as distance increases.

Cyberthreats are also becoming a critical issue in the interconnected digital business economy, making the public internet an increasing security challenge. Unfortunately, long-distance MPLS connections, while traditionally more secure, are expensive and also subject to the laws of physics.

Placing key IT and networking functions at the digital edge, close to dense populations of users and partners and high concentrations of applications and data, is a good first step to mitigating the risks associated with both latency and security.

Localise and optimise traffic

 The following steps will enable you take advantage of proximate, direct, private interconnection for better performance, tighter security and aggregated network connections among peering partners:

Strategically place digital edge nodes in global colocation facilities to establish a geographic hub based on dense user populations and business opportunities.

These edge nodes are communications hubs that act as the intersection points (shortening the distance) for multiple cloud, network, ISP and other service providers. The hubs can be interconnected in a global and distributed mesh configuration (see diagram below).

Enterprises are moving external facing services out to these hubs, such as VPN termination or direct internet connectivity to corporate data centers, which is not the same as internet offload, where they are removing circuits.

Solve the last-mile connectivity problem via a metro Ethernet instead of MPLS or by leveraging internet/VPN locally as an access alternative for mobile workers to get to the hub.

This enables LAN-like throughput, rather than costly and lower-performing, latency-sensitive connections. Concentrated urban mobile users experience significant network performance improvement over shorter distances.

Use intra-colocation (cross connect) networking to enable new connectivity paradigms through more cost-effective business peering and direct cloud connection.

Implement competitive inter-colocation access to reduce WAN costs, reduce latency, improve availability and provide new topology offerings regionally or globally.

You can use the combination of intra- and inter-colocation networking to derive the same benefits in performance and efficiency enjoyed by cloud service providers.

Apply traffic management across redundant connections to load balance traffic across your network fabric for greater performance, as well as allow for high availability and faster routing convergence in the event of a network failure.

Reduce the risk of cyberthreats and data breaches by limiting the exposed network to the last mile, and place security controls at the edge along with your IT and network infrastructure. This extends your security perimeter closer to where it is most needed.

Localising and Optimising Network Traffic Design Pattern

By applying a localised and optimised strategy, not only do you reduce interconnection latency at the edge from 20 milliseconds (over 20 hops) to 1 millisecond (over 1 hop), you shrink your long-distance interconnections cost from $100 per Mbps to under $10 per Mbps.

Use case

One use case where traffic localisation and optimisation is proving successful is with PaaS and SaaS providers and their enterprise customers.

In a traditional PaaS or SaaS offering, enterprises consume the services over the internet without any special attention paid to the number of hops or latency between them and the PaaS/SaaS service provider.

To improve their customers' QoE, service providers are quickly realising how important the digital edge is.

Instead of deploying in one region, they are offering networking and routing services in many geographic regions that are as easily accessible as the rest of features that make PaaS and SaaS platforms extremely compelling.

This translates to the network becoming an integral part of the PaaS and SaaS offerings, with service providers bundling premium routing and private connectivity as a part of those services.

For example, a rapidly growing national healthcare firm was in dire need of a fast, high-performance and cost-effective way to enable network connectivity between network service providers (NSPs) and cloud service providers for its acquisitions activity.

Leveraging an Interconnection Oriented Architecture (IOA) strategy deployed on Platform Equinix, the company enabled localised connections to the NSPs and CSPs via the Equinix Performance Hub and Cloud Exchange solutions and significantly reduced its access costs.

It also was able to meet its business integration timelines by optimising traffic between two new businesses, increasing the QoE of its users accessing applications.

This connectivity model can now be applied in a distributed fashion across several locations with a positive national and global impact.

Article by Olu Rowaiye - Erik Hensarling, Equinix blog network