Network slicing is not exactly a new idea. As network virtualization implementation has risen sharply in the last few years, organizations have included network slicing in the design and implementation of their virtualized infrastructure. Defining network slicing, particularly concerning its importance to 5G networks, is important as a precursor to any discussion of its capabilities, techniques for realization, and management.
What is Network Slicing?
Network slicing allows multiple “logical” networks to share a single physical infrastructure. Network slicing allows a single physical network to virtually look like dedicated, separate networks, each with its own capabilities from the perspective of bandwidth, latency, resiliency, privacy, security, and isolation. This allows a network operator to partition its network resources to allow for different users or “tenants” to multiplex bandwidth and services over a single physical infrastructure.
Network adaptability goes along with network slicing and adds to its benefits. Network slicing enables network adaptability by carving physical resources into individual networks, each controlled individually. Each of these “virtual” networks can be modified at a moment’s notice to change their path through physical networking elements, change redundancy characteristics, and/or per-tenant performance requirements.
These adaptations can occur dynamically through numerous mechanisms, including traditional network management, SDN controllers, or other standard orchestration techniques like OpenStack. The adaptability offered through virtual network slices is key to making network slicing an effective technique to support varied service level agreements, application requirements, and security profiles.
Virtualization and 5G
The virtualization of networking infrastructure has been going on for years via techniques like IP VPNs (IPsec, SSL, MPLS-BGP), MPLS switching, VLANs, or any other tunneling technology that separates traffic based on user, application, or service requirement. While important in any of these virtual network overlays, network slicing is interesting in the case of 5G.
This isn’t because the virtualized network infrastructure is inherently different than in previous techniques, but rather due to the breadth and varying applications that will be using this infrastructure for transport and the requirements inherent in these varied services over 5G.
Network slicing can be implemented for various reasons. The most basic is to provide a dedicated set of network resources that are guaranteed to a customer. These resources may be as simple as isolation from others to more sophisticated divisions for guaranteed throughput per customer or guaranteed uptime.
5G technology and services bring built-in improvements that will enable new use cases not currently achievable over a mobile network. For example, home Internet access, IoT, and mobile broadband all have very different requirements from a network perspective that need to be considered. For simple Internet access or mobile broadband, customers will demand guaranteed uptime and a guaranteed amount of throughput with few other demands. For IoT, on the other hand, network providers will need network slices potentially requiring lower throughput but with extremely low latency for sensor-to-sensor communication at the network edge, supporting millions of endpoints or sensor locations.
Other application use cases may be transactional in nature but are also time-dependent where a tenant wants to “dial-up” or “dial down” their network resources based on time of day, product launches, events, or other dynamic, business-related issues. Effective network slicing makes all these cases possible.
Challenges and Requirements
Monetization is the clear advantage from an operator’s vantage point. The ability to provide managed network slices for the previously described and other, unimaginable use cases allows carriers to charge on a per-tenant, per-user, per-application, per-SLA, per-uptime basis. This gives the flexibility to offer services that selling a “physical port” would never make possible.
Managing a network at this level of granularity is the next hurdle to overcome. Network operators simply can’t manage the massive scale and breadth of their infrastructure on a per-device basis. Network slices are end-to-end constructs that span the mobile and physical portions of the network connecting wireless node to wireless node, to physical infrastructure, and to the broader Internet. This essentially is a fully developed networking mesh, providing any-to-any connectivity at all times.
This level of connectivity requires programmatic control of all these resources seamlessly, regardless of the physical location of the network equipment. Administrators require a holistic view into the entire network from a logically single location to “see everything” and change the entire behavior of the network from a single touch point. Countless techniques and technologies exist and are being developed to enable the dynamic nature required to manage a virtually defined, network-sliced environment. This encompasses any virtualized network, be it cloud datacenters, SDN, or NFV.
Infrastructure Capabilities
The requirements for managing virtualized infrastructure span all of these “technologies.” In addition, it’s important to note that slices need to be adequately monitored and secured. That requires that the slices can be offered adequate security at the network edge and be monitored from any location in the network for troubleshooting and recording purposes.
Effective implementation at the right cost point is the most important factor in
network slicing. How can network operators support 5G services with effective network slicing? It’s a beautiful vision to consider the level of dynamism virtualized network slices promise, but to bring this to reality is a different story. If slices require a level of throughput, how is this guaranteed? How can we promise a certain measured latency? How is it measured? How can we promise dynamic routing around network failures and constantly updated security policies? This all comes down to the capabilities of the infrastructure itself.
In the conclusion of this article, we will look at how to make 5G a reality using data plane acceleration based on FPGA-based SmartNICs.
