The ability to “virtualize” is critical for operators evolving toward ThinkingNetworks™.
Network Function Virtualization (NFV) implementations and virtualization of the Evolved Packet Core (EPC), as well as systems outside the EPC (e.g., billing), can significantly reduce network costs and help operators become more efficient in matching resources to network and service demands. NFV gives service providers the ability to elastically assign compute and storage resources through a software-only approach.
Assigning resources only where needed is important when it comes to Diameter network elements, especially the Diameter Signaling Router (DSR) and Policy Server (PCRF), which have to be “cloud ready” in order to successfully control LTE EPC functions.
Virtualization will essentially partition the resources of a hardware platform into unique “virtual machines.” These virtual machines replicate standalone functions currently supported on separate hardware. If more compute resources are needed to support a Diameter function, any available hardware can be chosen and a new instance of the virtual machine created by the hypervisor.
The same hardware could also be used to support instances of a policy function at the same time, if enough compute resources exist. It’s that ability to dynamically allocate additional compute and storage resources when needed – using a common pool of hardware – that makes virtualization so important.
The move to NFV implementations will mean operators expand virtual functions to support multiple regions, or extend functions to other partners as part of their cloud offerings. For example, MVNOs or multinational operators looking to put their packet core into the cloud can manage all services and countries through one implementation.
A Closer Look At NFV
At Tekelec, we have developed the Orchestrator, which determines through analytics and other inputs when additional resources are needed. The Orchestrator instructs the Virtualizer (the hypervisor) to create another instance of a function. The Virtualizer creates additional resources, and other instances of the DSR or PCRF per the Orchestrator’s instructions.
Since the Orchestrator is the function that determines what, when, and where additional compute and storage resources are needed, it goes into action when traffic levels in the Diameter network increase. It identifies the rise in traffic and determines if additional DSR or PCRF resources are needed to support the rise in traffic.
Once the Virtualizer implements the new instance of DSR or PCRF, the IP Flow Manager directs IP flows to the new instance of the function (since routing tables will not reflect newly configured hardware).
The communication and cooperation among these components gives networks the ability to expand and contract based on real-time traffic conditions. That capability will become invaluable to operators trying to balance the need for innovative services with the need to maintain network performance levels – both essential to the customer experience.