By Daniel Kwan, Vice President and General Manager of Asia Pacific excluding Japan, Sonus
Within a few years, analysts predict that enterprise applications will be in the Cloud more than out of it, passing the halfway mark—that is, more than half of all business apps will be in the Cloud—somewhere around 2020. In Asia Pacific, IDC has predicted that in the next three to five years, 33 percent of the region’s enterprises will implement a Cloud-first strategy.
For communications services, the timetable to the Cloud is a bit complicated. Building out a virtualized communications network is a complex process that involves a lot of unique functions and capabilities, particularly around real-time communications and quality of service (QoS). The need for service providers to virtualize their networks is gaining urgency as the market for hosted Cloud services and mobile apps expands. In order to compete effectively in those markets, service providers will move to a virtualized, Cloud-based infrastructure to drive down costs in the face of rising traffic and, more importantly, deploy new revenue-generating services quickly.
By now, most service providers realize the inherent advantages that Network Function Virtualization (NFV) can offer in terms of capex reduction and scalability. Fewer realize that virtualized network functions (VNFs) also serve as a foundation for service orchestration and service automation—technologies that can dramatically accelerate and simplify service creation, in some cases reducing the time to market for new services from months to minutes. It’s this capability that will allow service providers to compete effectively with over-the-top (OTT) players for a larger share of the mobile app and Cloud services market. This prevents service providers from being relegated to the bit-pipe providers of the 5G future.
Virtualization, Automation and Orchestration
Virtualization refers to the concept of replacing physical network appliances with virtual instances that can be deployed on commercial-off-the-shelf (COTS) servers. A service provider network is significantly different than an enterprise data center, with specific network elements to support real-time communications such as session border controllers (SBCs) and IP Multimedia Subsystem (IMS) gateways. Virtualizing these elements is more complex than simply loading generic software onto an Intel-based server and virtualization alone does not make an application optimized for the Cloud.
Service automation is the ability to automatically provision services on network elements. In the legacy network architecture, service provisioning is a manual and time-consuming process that requires each appliance to be individually configured for each new service and service update. For this reason, rolling out a new service that involves tens or hundreds of different elements can take months of configuring and testing before it’s ready for production. In a virtual, Cloud-based environment, service providers only need to create a template for a master configuration and then use that template to replicate the service provisioning process across virtual instances. What previously took weeks now can be done in hours.
Service orchestration provides coordinated management and control of network services across multiple network domains that are responsible for connectivity. Once a service provider automates a single service, they can stitch different services together, which is the orchestration part, to create new services. The company can even feed network intelligence into the orchestration engine to create new services based on individual behaviors or unique characteristics. With service orchestration and automation, it becomes possible and profitable to create targeted services on demand.
The Road to Virtualization Starts Here
Transforming a legacy communications network into a virtual, Cloud-optimized network isn’t a walk in the park, it’s a journey. Every journey needs to start somewhere, preferably on a path to short-term ROI so that the network can experience the benefits of virtualization sooner. For mobile service providers, virtualizing the mobile core can be a good starting point. When building out a virtual mobile core, it really needs to be in parallel with the legacy core to prevent disruption of critical services. At the edge of the network, the risk is lower and the short-term benefits can be greater. For this reason, many service providers may wish to start their journey with an element like their session border controllers.
A virtual, Cloud-optimized SBC offers immediate short-term gains through capex reduction and increased scalability. Service providers can deploy real-time communications services in new regions in minutes using virtualized SBCs, compared to physical SBCs that may take weeks to set up as these include cabling, installation, as well as housing and energy costs. They will often deploy virtualized SBCs to expand, rather than replace, their legacy SBC architecture, which is a good idea so long as the physical and virtual platforms deliver the same features and user experience.
As voice and video call volumes increase, SBCs need to scale up to handle the increase in SIP traffic as well as the increase in encryption (e.g., IPsec, SRTP) and media transcoding. As service providers become Cloud service providers, virtual, Cloud-optimized SBCs will also allow them to easily and securely shift SBC resources to support dynamic traffic requirements.
In a very real sense, SBCs are service providers’ gateway to the Cloud. Virtualizing those gateways opens up the network to more opportunities while increasing security as well as revenue.
