For communication service providers, the move to a cloud-native environment has impact across the board.
Larry Rau, the director of architecture and infrastructure for Verizon Labs, says that a cloud-native environment is one in which commodity servers are managed and controlled as a group and host applications offered that are part of a collection of small, or micro, services. The services have some level of virtualization and therefore are isolated from hardware and other software components.
Rau elaborated on Verizon's approach to cloud and cloud-native architectures in this Q&A with Telco Transformation.
Telco Transformation: Operators want cloud-native VNFs, BSS and OSS systems. What does this mean for operators, their IT support systems and their cloud infrastructure?
Larry Rau: The important shift in this area is a move to a software-defined everything. All functions for operating the network move away from dedicated hardware, or specific hardware, to the use of commodity servers -- uniform servers, so to speak -- with all the function and control happening via software modules. The result is a more consistent set of hardware resources, leveraging industry-wide economies of scale for that hardware. Further, software modules are quicker to enhance, replace, and manage. This coincides with adopting a software infrastructure that allows for efficient management of your hardware resources -- or the so-called cloud native.
TT: What does this mean for the companies they are working with?
LR: Many traditional suppliers of both infrastructure and service solutions are adapting or should be adapting. This shift is an important inflection point for the industry. It's a time for new competition among existing suppliers as well as newcomers, which stimulates the industry to improve.
TT: What is the relationship between cloud native and popular new technologies such as SDN, NFV, containers and virtualization?
LR: While this can be subjective depending on how each of these is defined, the common thread among them is the heavy focus on software.
TT: Like any other new technology, cloud native is defined a bit differently by different companies. What is your definition?
LR: We look at cloud native as a move to a uniform set of commodity hardware servers that are managed and controlled as a large group and that host sets of applications designed as a set of inter-operating small services (microservices). These applications will have some level of virtualization to allow for independence from hardware and other running software components -- typically, and specifically in our case, using container technology as that unit of virtualization.
TT: Since cloud native is horizontal, what precautions have to be taken to ensure that things happen correctly over time and distance?
LR: In many ways a cloud-native solution does not change the need to consider critical performance attributes for some types of functions. Verizon has a lot of experience planning, developing and operating a global network and data centers that meet these performance and reliability requirements. Grouping functions into different data centers in a way that makes sense allows us to take advantage of the scale benefits of a cloud-native architecture and is fundamental to our planning.
TT: How does cloud native impact the actual programming? In other words, how does the actual app that is being developed (as opposed to the process) change?
LR: Developing applications for a cloud-native environment is critical to fully leverage the cloud-native type of architecture applications needed for a horizontal type of scale and to make use of virtual interfaces to avoid being tied to specific hardware. Fortunately, many existing software practices and tools lend themselves to this type of programming and environment. Of course, newly developed applications have an advantage because applications being migrated from a more rigid infrastructure could have some challenges to work through.
TT: What steps is your organization taking going fully cloud native?
LR: Verizon has publicly come out in support of many initiatives adopting SDN/NFV, virtualization, and data center scale computing. Internally, many development activities have been going on to move to a foundation that adopts a cloud-native architecture. Publicly we have demonstrated our use of Apache Mesos and containers in general, as well as OpenStack and virtualization, and use of commodity hardware to operate these platforms.
TT: What advice would you give a telecom company beginning to create applications using cloud native?
LR: The sooner critical mass can be achieved in a cloud-native environment, the sooner the full benefits of the environment can be achieved such as operational simplification, leveraging capital assets efficiently, and the velocity of adding new or enhanced functionality.
— Carl Weinschenk, Contributing Editor, Telco Transformation