|Cloud Scale Multi-terabit Transport – Scale Up Instantly|
September 20, 2017
By Dan Parsons
Sr. Manager, Architecture Marketing
Cloud, cloud, cloud, cloud…it seems that every other word in our industry is “cloud.” Rightly so, because it has become the center of the new era of the digital communication and information technology (IT) landscape for business. Almost everything we do with our smart devices involves the cloud, and connectivity is required for many of the apps we run on our tablets, laptops and smart phones.
The cloud is essentially the infrastructure of the World Wide Web transformed to an amalgamated compute, storage and network infrastructure. The compute and storage resources are virtualized and hosted on servers in data centers (DCs) that are interconnected all over the world through an optical transport network.
User bandwidth growth is tracking Nielsen’s Law at about 50 percent per year, and the number of connected devices is rapidly moving toward an estimate of 50 billion by 2025. As end-user traffic grows so does the traffic between data centers, and data center interconnect (DCI) capacity is growing at almost 50 percent per year. As a result, cloud scale transport is imperative for optical infrastructure.
What Is Cloud Scale Transport?
Fundamentally, it is the same principles that are used within the DC applied to the transport network –capacity and performance that are instantly scalable from a virtual infinite pool of resources, with standards-based open software and networking. Using these principles, the cloud is resilient, automated, offers easy point-and-click operation, is enabled by simple rack-and-stack equipment and is programmable to host any application.
Figure 1: Principles of Cloud Scale Transport
These principles are applied in part to traditional telecom transport equipment. However, without an underlying core photonic technology and architecture, traditional transport equipment cannot scale with the cloud.
Consider just the cloud scale principle of scaling capacity. Within the DC, performance can be scaled up on a single server and/or scaled out across multiple servers by instantly accessing a virtual infinite pool of resources. It appears infinite because compute and storage resources can be returned to the pool and used again, or in cloud terms, scaled up/down and out/in. However, for traditional optical transport equipment, instant scale up is limited to a single wavelength and not the entire spectrum if no other wavelengths are available and ready for service.
The value of cloud scale transport can be seen in subsea applications, where only a few fiber pairs are typically available and distances can be more than 10,000 kilometers. In addition, subsea capacity demand increases at about 50 percent per year, driven primarily by content service providers. A subsea communication service provider (CSP) has two options to address this issue: either to increase the capacity of the current optical channels on the transport network or to add more channels. For subsea transport, increasing channel capacity may not be viable – for example, if the capacity-reach performance limits of the current technology have been reached and 50 percent more capacity is required, the only option then is to add more optical channels or wavelengths. The CSP faces the dilemma of how many channels to add – what will result in the best return on investment (ROI)? Subsea wavelengths are not cheap. Unless the CSP has idle optical channels on its multi-terabit transport network, the following series of events must take place to scale up the network with the new optical capacity:
|Planning (Engineering)||+ 1/2 week|
|Ordering hardware (Procurement)||+ 1/2 week|
|Delivery||+ 10 weeks|
|Installation and commissioning (Operations)||+ 2 weeks|
In this typical deployment example, the new capacity is only turned up after more than two months involving three different departments. Although the cloud drove the demand, the traditional subsea transport network could not scale with the cloud. The real impact was that about three weeks of operational costs were incurred and two months of revenue was lost. In the age of the cloud, this is unacceptable. Clearly, cloud scale transport is more than just big fat pipes – it’s the ability to scale up optical capacity instantly and economically from an infinite pool of resources, just like inside the DC.
Consider the cloud scale Infinera Intelligent Transport Network with the embedded fourth-generation Infinite Capacity Engine (ICE4) and the Instant Bandwidth solution for the subsea application. A flexible, pre-qualified, multi-terabit-per-second pool of capacity is available that can be accessed instantly for scaling up and down services with the click of a mouse. The ramifications of this are enormous. Not only are almost two months of the traditional deployment cycle eliminated, but the CSP can respond to business opportunities quickly, realize faster time to revenue and offer new types of service such as on-demand wavelengths. With cloud scale transport, the subsea CSP can offer competitive and differentiated services and realize better ROI, leading to higher brand value.
The ability to scale up transport capacity at cloud scale is not only obvious in its value but also necessary. Leveraging the combined cloud principles of scaling up capacity, scaling out the network, open software and networking and an infinite pool of virtual capacity is possible with key technologies like ICE4, intelligent transport architectures and application software. In future blogs on cloud scale transport, we’ll consider the other three key principles – stay tuned!
For more information, please contact us.
Webinar: Cloud Scale Optical Network Questions? Get the Answers
Web page: Infinite Capacity Engine
Blog: The Five Steps to Cognitive Networking (Part 1)
Blog: The Five Steps to Cognitive Networking (Part 2)