In my previous post I explored how deployments can mix thick and thin vCPEs to accommodate factors such as enterprise size, number of sites, connectivity constraints, etc. Two additional dimensions also play a key role in determining optimal deployment architecture:
1. The nature of services being delivered
2. The integration between services and the operator’s OSS/BSS environment
The figure below shows examples of enterprise sites using thin, medium, and thick vCPEs to support diverse services.
With a thin vCPE, VNFs are deployed in the cloud, reducing operational expenses and enabling operators to adopt a web-like model where customers can add and modify services through a web portal. In this case, services such as analytics & reporting and service chaining between cloud hosted VNFs (e.g. firewalling, content filtering, etc.) are well served. However, this has a major impact on existing OSS/BSS infrastructure.
With a thick vCPE, a more powerful device is deployed at the customer site to run some functions locally. This type of configuration can be required to support services such as subscriber/application-based QOS and traffic prioritization, WAN acceleration, and local routing. This approach provides less agility for adding and modifying services, but it also requires less adaptation to OSS/BSS suites as some services continue to run at the customer premise.
In the next post, I will explore how Layer 7 visibility can be leveraged to provide value-added vCPE services.
Francois de Repentigny, guest blogger, Level Up Marketing