To successfully "cloudify" the network, it's important to recognize how it differs from the data center.

Prayson Pate, CTO, Edge Cloud, ADVA

February 16, 2017

4 Min Read
Cloudifying the Communications Network

Operators want to take advantage of cloudification, but the communications network is not the data center. So, how can operators bridge the gap between the network and the cloud? How can they make a transition from current operational methods and realize the benefits of cloud technologies? Let's take a look at the concept of cloudification and how it applies to the communications network.

What do we mean by cloudification?
People often misinterpret cloudification to simply mean virtualization, but it's actually much broader than that. Here's what's included in cloudification:

  • Disaggregation -- Network appliances (such as routers and firewalls) are broken into separate hardware and software components that then may be acquired from separate suppliers or open source projects.

  • White box -- Open-third party servers and switches are used to provide a low-cost and programmable infrastructure.

  • Components and APIs -- Rapid construction of services is enabled by assembling software components using open APIs.

  • New development methods -- Agile development and DevOps increase velocity.

  • New commercial models -- Embracing new ways of working with customers and suppliers including pay-as-you-go and shared success.

Mapping from the cloud to the communications network
Cloudification stems from data centers and cloud-centric companies. The data center has a number of characteristics that make it suitable for the principles that contributed to the growth of the cloud:

  • Homogeneous -- All servers and storage arrays are very similar, with few differences to account for.

  • Green field -- Deployed equipment is all relatively new and feature-rich.

  • Geographically colocated -- Colocation makes VM mobility and system reliability a reality.

  • Cheap bandwidth -- Distributing compute and storage resources in a data center is easy when high bandwidth and low latency are the rule.

  • Low latency -- The adjacency of equipment keeps communications latencies very low.

  • Unreliable elements -- Cloud resiliency is built into the system and software, not the physical elements themselves.

However, the communications network is not the data center. Its characteristics include:

  • Heterogeneous -- A wide array of equipment with different capabilities is deployed.

  • Legacy -- Equipment may have been deployed 20 or more years ago. Capabilities for new features are limited.

  • Geographically dispersed -- No assumption of adjacency can help with the optimal pooling of resources.

  • Expensive bandwidth -- VM mobility over a 1Mbit/s access link is not practical.

  • High latency -- Network elements may be dispersed globally.

  • Lack of trust -- Public networks are frequently considered to be untrusted domains.

  • Long service lifetimes -- Connectivity services are provisioned for weeks, months or even years.

  • Reliable elements -- Each network element is individually reliable, sometimes with redundant components.

In order to embrace cloudification, operators must acknowledge and address the differences between the data center and the communications network, listed above and summarized below.

Figure 1:

Realizing the benefits of cloudification
As you begin the process of cloudifying your communications network, here's a list of key considerations the leading operators have discovered as they have sought to realize the benefits of the cloud:

  • Look for partners with a history of providing software-based solutions to operators.

  • Build on a base of foundational cloud technologies, such as Linux, KVM and OpenStack.

  • Work with companies with a large ecosystem of partners to deliver maximum flexibility.

  • Address deployment and scalability problems with software features like software-based Carrier Ethernet, zero-touch provisioning capabilities and embedded cloud functionality.

  • Make the business case work for edge deployments by optimizing the performance of low-cost standard servers.

  • Create a unified architecture composed of software elements that can be disaggregated to work in a variety of environments.

  • Implement additional security controls for connecting customers with a data center in a protected and reliable way.

  • Instantiate workloads on demand in real time with an automatically provisioned connectivity network preferably through open interfaces and SDN control.

It's time to leverage NFV and embrace the principles of the cloud in your network. And you don't have to do it alone. There are partners available with the experience, products and services to make NFV easy and cloudification real.

— Prayson Pate, Chief Technology Officer, Ensemble, ADVA Optical Networking

Read more about:

Europe

About the Author(s)

Prayson Pate

CTO, Edge Cloud, ADVA

Prayson is passionate about helping the telecom industry go virtual. He believes network functions virtualization (NFV) is the key to profitability and evangelizes about essential benefits achieved by using big data analytics and pure-play NFV orchestration to optimize a carrier-class virtual networking environment. Prayson speaks about NFV and software-defined networking at industry events, posts regularly to industry blogs and his own site www.praysonpate.com, and contributes articles for publication in industry media. As CTO at Overture he is focused on making NFV real and profitable for communication service providers (CSPs.) Prayson is an accomplished technologist who began his career in 1983 at FiberLAN. He also worked for Bell Northern Research and Larscom before co-founding Overture in 2000.

Subscribe and receive the latest news from the industry.
Join 62,000+ members. Yes it's completely free.

You May Also Like