Nationalistic trade wars aside, the world’s economy is truly global, and globally distributed enterprises are aggressively expanding their business into more countries. Growth is especially brisk in the Asia Pacific region and China in particular. To establish their facilities, companies need reliable and high-performance network connectivity to global data centers both in-region and out-of-region.
Companies that are accustomed to the reliability and affordability of high-performance connectivity in North America and Europe might be challenged by WAN issues as they venture into less developed regions. Internet infrastructure is often less developed, which can lead to problems such as packet loss over the last mile. At the same time, MPLS circuits can be quite costly and take many months to install.
The two top WAN issues for many global enterprises are the quality of last-mile infrastructure to remote locations and the high latency of global connectivity.
- Last-mile considerations – Internet last-mile connections in developing countries are often less reliable than in North America or Europe. This may be due to poor physical infrastructure or an oversubscribed connection shared by many businesses (such as with cable or ADSL). The situation is improving with countries upgrading their infrastructure. Still, a reliable Internet link may be unavailable, requiring the purchase of MPLS last mile.
- Global connectivity considerations – When delivering applications across long distances, latency and packet loss – not bandwidth – will determine application performance. This established fact becomes critical for Asia Pacific connections. The long distances and poor Internet peering between the Asia Pacific, North America, and Europe exacerbate latency. Infrastructure problems and oversubscription can increase packet loss.
There are additional considerations for WAN connectivity as well, such as a high percentage of application traffic now going to the cloud, and applications such as voice and video conferencing that require high quality of service. Data and application security also are critically important.
Top WAN Issues are Addressed by a Cloud Network Architecture
There is now an alternative to the traditional approach of using global MPLS services to build a WAN that can address all the issues of last mile connections, latency and packet loss across distances, cloud usage, quality of service, and network security. The new model to meet the networking needs of globally distributed enterprises is a cloud network, also known as software-defined WAN (SD-WAN) as a service.
Cloud networks revolutionize global connectivity. Using software, commodity hardware, and excess capacity within global carrier backbones, cloud networks provide affordable SLA-backed connectivity at a global scale. Cloud networks deploy edge devices to combine last mile transports, such as fiber, cable, xDSL, and 4G/LTE, to reach a regional point-of-presence (PoP). From the regional PoP, traffic is routed globally to the PoP closest to the destination using tier-1 and SLA-backed global carriers. By keeping the traffic on the same carrier backbone worldwide, packet loss is minimized, and latency can be guaranteed between global locations.
What’s more, a range of optimizations can be applied to get even better performance across both the “middle mile” as well as the last mile. The middle mile – i.e., the global backbone of the private cloud network – is typically engineered to have zero packet loss. As for the last mile, some cloud network providers are able to apply packet loss mitigation techniques to regenerate lost packets traversing this portion of the network, thus ensuring high quality service.
This global networking model extends to cloud services as well. Traffic to SaaS applications like Salesforce.com and Office 365, or to cloud data centers such as Amazon AWS and Microsoft Azure, will exit at the PoP closest to these services—in many cases within the same data center hosting both the PoP and the cloud service instance. This is a dramatic improvement over the unpredictable public Internet and a significant cost saving versus the expensive MPLS option.
Additional features can provide even more benefits to using a cloud network. For instance, a known challenge to having a WAN presence in China is “the Great Firewall of China.” Due to regulatory oversight, traffic leaving the country must be inspected by a central firewall which enforces Chinese regulations regarding the use of Internet and cloud services. As a result of this policy, global Internet-based connectivity from China exhibits high packet loss and high latency. However, a cloud network with a government-approved link can allow traffic exiting China to enjoy consistent low latency and zero packet loss as traffic is optimally routed to all global enterprise locations and cloud applications.
A cloud network also has inherent network security, which enables customers to enforce centralized security policies on WAN and Internet traffic without distributed firewalls. This reduces the cost and complexity of IT services in the remote locations.
As the global economy becomes even more intertwined, distributed enterprises will be reaching further into new regions where they can produce and sell their goods and services. This global marketplace must have the support of reliable and high-performance network capabilities to sustain and grow the businesses. A global cloud network facilitates the reach and performance these enterprises need, now and into the future.
You may also be interested in: How to Deliver Reliable, High-Performance WANs into Asia Pacific and China