Goals and Challenges for CSPs in Network Automation

In the last few years, communication service providers (CSPs) are going through major infrastructure transitions to deal with increasing bandwidth demands from subscribers. A transition includes integrating cloud model into a backhaul of telecom network, as central or edge cloud. This transition has a significant impact on the whole IT technology business and the overall IT ecosystem. Now, with this, we will see how IT companies are playing a crucial role in building backhaul technologies for telecom network and there we have a big shift in how IT pros are building software technologies and how subscribers will access services.

CSPs are trying to lower their internal (operations and resources) costs to keep service costs to enterprise users lower, improve service delivery, and ensure and enhance the security of their offerings to subscribers. At this point, IT companies come into the picture to help CSP to transform hardware-based network to virtualized, i.e. software-defined network to bring agility, automation into a network with lowering expenses on resources (CAPEX) and operations (OPEX). 

Apart from the expenditure part, network automation for deployment and management of digital nodes will be the key to make this move successful. IT technology vendors are service providers are going to play a major role in this.

Network automation enablement with cloud

The reason for this move for telco cloud is due to the fact that cloud service providers are constantly creeping into network infrastructure to deliver end-to-end services. This is nothing, but we can call it as an amalgamation of CSP networks with cloud providers. Moreover, this opens up various possibilities for CSPs in terms of handling dynamic scaling of network resources and cloud applications, and automation of network operations to provide service velocity, fault rectification, and security.

In this context, in December 2018, Futuriom research firm released a report on CSP network automation to get insights from industry experts. A survey was conducted and gathered data from 130 professionals including telecom service providers and web-scale cloud operators. The focus of this survey was to understand key goals, challenges and components of network automation for CSPs. The report shows a comprehensive perspective on how automation can be helpful in upcoming years and an approach by service providers and cloud service providers.

Why are the expectations of CSPs from network automation?

If we look at the current network that is managed by CSPs, it is legacy network constituents of telco hardware and OSS/BSS on top of that which is very difficult to deploy each node and services on top of that manually. It will also be a huge task ahead in front of CSP to transition network to a virtualized one. The complexity of virtualized network is also in the testing phase due to the complex nature of virtualized deployment of operational elements within the network. So, what is in hand of CSPs – a vast telecom network which is going to transform into software-driven containing innovation provided with software-defined networking (SDN) and network functions virtualization (NFV). Also, on the road to 5G, new architecture like network slicing with edge computing will be adding more critical architectural ecosystem.

Due to all this end-to-end automation, become a very critical requirement for CSPs for driving efficiency by saving time for configuration and deployment along with cost as well by eliminating hardware and operational expenditure.

As per the report, there can be five expectations if automation comes into picture for CSPs:

In the survey, Futuriom asked about the top three goals out of above expectations to participants. Faster service delivery for acceleration in revenue and support for dynamic services came on top of the selection. Network security with intelligence was the next priority for participants.

Therefore, it is clear that CSPs require network automation to focus more on service velocity and faster delivery along with secure network operations.

In any software driven environment, automation of operations and resource deployment mainly has an impact on cloud-native technologies. Containers, container orchestration engines, and microservices architecture along with central management platforms have been introduced by the public as well as private cloud vendors to fuel automation. As the CSP network evolved with the software-centric approach, the cloud-native technologies have been merged to the different tiers in the telecom network.

CSP relies on the network at highest scales where the realization of end-to-end automation is a hard thing. The main goal for CSPs can be to simplification of network design for integration of emerging cloud-native technologies to achieve automation. But there are certain challenges comes into the picture like:

A question was asked in a survey to find the most common challenges based on the above list. Investment in resources for implementation of network automation for staff and money is a topmost challenge according to the survey respondents. Next barrier came out is roadblocks in implementing new technologies at scale.

Therefore, it is highlighted to vendors and CSPs to the reality of lack of implementation assistance and tools in network automation.

SDN and NFV technologies are playing a key role in virtualizing network elements and abstracting software part of network hardware to have a central dashboard. The virtualized software sits on top of the COTS hardware at data center stacks within the network, which is again further used to gather telemetry data and analytics information. Furthermore, this pushes for even more intelligent automation that can be called as intent based networking. Most of CSPs are focusing to have the ability to just pass on the intent at central dashboard and network is deployed and managed accordingly.

The software approach in network automation enabled with SDN, NFV allows CSPs to gather telemetry, analysis, and monitoring data, which can be additionally used for security, purpose and allow CSP to have a network with intelligence fault remediation capabilities.