Implementing Policy Control as a Virtual Network Function: Challenges and Considerations
An Industry Whitepaper
Traditionally, the processing demands of policy control (e.g., stateful packet processing, complex decision-making, etc.) required proprietary hardware solutions, but technology advances mean that virtualization now, or at the very least soon, provides an alternative.
Transitioning from a purpose-built, proprietary hardware component – one in which a vendor likely controls every aspect – to a virtualized COTS model in which performance is dependent on clock speed and available cores, and in which drivers vary by hardware manufacturer is a formidable challenge.
Vendors who embark on this transition face a number of considerations and must overcome many challenges in order to preserve network policy control functionality in a virtualized environment.
By understanding these considerations and challenges, communications service providers gain an informed position from which they can effectively evaluate alternatives.
To explore these topics, this paper asks and answers the questions:
- How can a solution maximize the performance of each individual core?
- How can a solution maximize the performance of the overall system (i.e., of all the cores working together)?
- How can a solution effectively coordinate aggregate functions across many cores?