Analyst Angle: Virtualization or containerization for 5G?
It’s anticipated that 5G networks and services will begin to be operational by approximately 2020, but the preparations and investments in the foundational technologies are still being tackled. Virtualization and containerization are both networking architectures that use common, off-the-shelf (COTS) hardware in order to achieve greater scale and economics than purpose-built hardware. Virtualization uses a common software infrastructure to create and manage virtual machine (VM) lifecycles upon which network functions and applications will reside. Containers, on the other hand, represent a full-software stack (operating system through application) that can be quickly instantiated as capacity requirements change.
5G standards will likely include core network topologies that are akin to what many of the leading over-the-top (OTT) service providers are using today. These new specifications will allow traditional network operators to quickly define, test, and deploy new services for a vast array of devices and services, some with nearly instantaneous response times and extremely high bandwidth (ex. autonomous cars), to services with extremely low bandwidth but very long battery life (ex. remote sensors). But the only way for all of this to work, from a scalability and economic perspective, is for 5G to be cloud-based.
A key element of cloud-enabled networks is “state-free” network functions and microservices. As people roam and rely on different access networks (ex. wifi or cellular) you can imagine that all the information coming from these devices will constantly be changing, thereby altering the state (whether the device is idle, transmitting, receiving, etc.) of the device within each functional entity. Since the state information is repeated in many different functional elements in the network, it would be more efficient to have the state information reside in a single database that is accessible to each of those applications. Due to this overload of shifting information, providers will need to offload that data from their applications, allowing them to scale a lot easier and require fewer computing and working resources.
Technologies like containerization and virtualization are enabling a drastic change to take place in network architecture, allowing traditional structures to be broken down into customizable elements that can be chained together programmatically to provide just the right level of connectivity, this is what we call ‘network slicing,’ with each element running on the architecture of its choice. This concept of network slicing is one of the key capabilities that will enable flexibility for core networks, as it allows multiple logical networks to be created on top of a common shared physical infrastructure. The greater elasticity brought about by network slicing will help to address the cost, efficiency, and flexibility requirements imposed by future 5G services.
Containerization vs virtualization
The ability to use low-cost hardware as the basis of networking infrastructure is a key foundation of 5G standards, as network operators begin to re-architect their core networks for virtualization and/or containerization, hoping that virtualization will be the beginning step to make core network functions more economical, scalable, adaptable, and open.
Operators will gradually adopt a cloud-based architecture for their core networks. A container-based network capitalizes on efficiencies of COTS-based hardware and the virtual machines, but what you lose with containerization is the flexibility of having common/ virtual machine platform. Eventually, core networks will resemble those of the large Web Scale Operators, like what Google, Amazon, Facebook and Netflix, run today. While containerization does allow for virtual functions to ride on top of a virtual machine, there may be some inefficiencies associated with it, but each server rides on a full stack of software.
Containerization is more efficient than virtualization in a lot of respects, but the ecosystem surrounding containers is still nascent. Managing and securing a vast amount of containers will be required as mobile computing workloads are migrated from more transitional equipment and applications riding atop of virtual machines. As they mature, container management efforts like Kubernetes, Apache Mesos, Docker Swarm will go a long way in making it easier and cost effective to create, manage and secure, large-scale container-based networks.
Given the relative immaturity of the standards and the ecosystem surrounding containers, many in the industry do not see containerization as being important yet and are too focused on virtualization, but it’s not the end game. Virtualization is more of an immediate step to cloud and containerization. However, it does have some inherent efficiencies that lead us to believe it will be the choice for operators to deploy 5G five to ten years down the road.
To support the wide range of performance requirements demanded by new business opportunities, multiple access technologies, a wide variety of services, and lots of new device types, the 5G core will be highly pliable. Minimizing cost for service providers and industries that depend on connectivity is a key part of the design for this flexible and dynamic core — enabling them to keep costs under control, while networks adapt as quickly as business models change. Network functions will no longer be located according to traditional vertical groupings in single network nodes, but will instead be distributed to provide connectivity where it is needed, the cloud will help to make this a reality. People expect that figuring out virtualization’s role will be an important part of this, but the focus for 5G will really be more around understanding the network topology.