Optimizing Streaming Efficiency with CDN Autoscaling: A Broadpeak and VMware Collaboration


Adopting a new application, service, or technology often results from a series of concomitant, favorable drivers.  

For instance, you might remember the explosion of mobile internet in the mid-2000s. That happened as the result of the iPhone launch and its revolutionary store of applications, the arrival of 3.5G1) offering consistent speeds of several Mbps, and the growing consumer appetite for internet access anytime, anywhere, on any device. 

On that note, this blog explores the favorable conditions for the adoption of virtualization in content delivery network (CDN) infrastructures. 

Why CDN infrastructures need to evolve 

So far, CDNs have been deployed as static appliances (mostly bare-metal servers) deployed once for all in internet exchange points (IXP), ISP points of presence (PoP), co-location facilities (so-called “colos”), or global CDN datacenters. For video service providers to scale content streaming for a medium-size country like France or the UK, a handful of CDN PoPs were necessary. 

Over the past few years, the overall number of CDN PoPs has been increasing, for several reasons: 

  • Streaming has exploded, with new OTT content providers successfully emerging in the market (e.g., Disney+, DAZN and Viaplay lately) requiring more (distributed) CDN capacity to ensure a consistent delivery to consumers over IP networks. 
  • Traditional video service providers have started transition their content delivery networks from legacy broadcast technologies2) to unicast HTTP-based adaptive bit rate streaming (ABR) in order to offer new services and reach all types of devices (smart TVs, 5G terminals, OTT set-top boxes, etc.), requiring more CDN capacity to optimize delivery. 
  • Many OTT content providers have been deploying their own CDN appliances in operators’ networks (e.g., Netflix Open Connect Appliances, Google Global Caches), to reduce their global CDN costs and improve streaming quality for customers. 

And this trend is only accelerating. The need for more distributed CDN capacity is greater than ever. 

However, increasing CDN decentralization is not viable without virtualization and orchestration. Deploying static CDN capacity on hundreds of telco edge nodes is just not cost-efficient, as streaming demand is constantly changing in time and location. Distributed CDN infrastructures have to be “elastic.” CDN workloads have to be deployed on-demand on multi-purpose underlying cloud/datacenter resources. 

The right time for CDN virtualization 

Fortunately, today there is a favorable situation for the virtualization of CDN infrastructures, resulting from the combination of a number of factors: 

  • The need for flexible, effective and cost-efficient CDN systems. 
  • The willingness of operators and service providers to harmonize their operating environments. After benefiting to IT, business support systems and more recently network functions, virtualization is now positively expanding to media application servers and service platforms. 
  • The emergence of cloud-native 5G networks and MEC3) platforms fostering distributed and fully virtualized architectures. 
  • The explosion of Kubernetes, which has emerged as an indispensable development, deployment, and orchestration platform. 

CDN virtualization revolutionizes the way CDNs are deployed and operated: 

  • Cache servers are containers running on multi-purpose telco edge cloud Kubernetes-based environments, and can be orchestrated along with other network functions. 
  • Cache instances are dynamically created and removed in a given edge PoP or MEC platform according to the demand in a specific region or area. 
  • Cache nodes scale out during spikes of streaming traffic and scale down once the peak is over. 

The result is an elastic CDN that follows streaming and caching demand in real time, guaranteeing the best quality of experience for consumers, especially for live content. Moreover, an elastic CDN mobilizes resources only when needed to save infrastructure costs and energy. 

Broadpeak and VMware collaboration on CDN elasticity 

Broadpeak and VMware CDN autoscaling architecture
Broadpeak and VMware CDN autoscaling architecture

Here is a video that explains the concepts and benefits of such a solution 


 The collaboration between VMware and Broadpeak brings all the aforementioned benefits of an elastic CDN to operators. Watch the video below to understand in details how. 

In the video, you can see how with just a few clicks, it is possible to: 

  • On-board Broadpeak’s container-based Open Dynamic and Elastic Advanced CDN. 
  • Leverage the Broadpeak Video Delivery Mediator and VMware Telco Cloud Automation APIs to automatically trigger up and down scaling of your cache servers based on real-time traffic conditions. 
  • Create or remove cache server replicas running on VMware Telco Cloud Platform in a matter of seconds. 
  • Zero-touch operate Broadpeak’s Advanced CDN without pre-assigned capacity. 
  • Deliver high-quality video streaming services with the confidence that underlying cloud resources are always utilized in a highly optimized manner. 

 We will be at VMExplore Europe, Nov. 7-10 in Barcelona. Let’s meet up!  


more blog

Internet Society

Connecting the Unconnected: An interview with Naveed Haq from the Internet Society

Youth’s watching habits : How is Gen Z consuming video?

Youth’s watching habits : How is Gen Z consuming video?