Why Virtualized Messaging is Vital Component for an Operator’s Messaging Infrastructure

Over the past few years, the global mobile messaging landscape has been undergoing a paradigm shift. Most notably, global person-to-person mobile messaging traffic has witnessed a significant decline, while over-the-top (OTT) traffic is on the rise.

To counter this, telecom operators are exploring new revenue streams and streamlining their operations. To illustrate, these players have chalked out plans to launch omni-channel messaging services, including new-age IP channels like chatbots. These initiatives are aimed at providing an integrated experience pertaining to application-to-person (A2P) messaging across traditional and new age channels.

Operators have turned their attention to A2P messaging, given that the global application-to-person (A2P) space is growing by leaps and bounds. This growth can be attributed to a number of factors-including low cost, high reliability and seamless accessibility. In fact, A2P messaging is already being deployed in a variety of use cases, such as authentication and verification, notification and alerts, promotions, payments and fund transfers, emergency broadcasts, etc.


The Role of Network Functions Virtualization in an Operator’s Messaging Infrastructure:

Are Operators Ready for NFV?

  • By end of 2018 66 per cent of SMS will be virtualized.(Heavy Reading)
  • This will be driven by reduced hardware-related operating expenditure
  • Virtualization ensures ease of scaling network capacity and enhanced operational efficiency

To leverage the full potential of A2P messaging, operators need to look at transitioning their messaging infrastructure as virtualized network functions (NFV). NFV can well be considered a key component, given the multiple benefits accrued. These largely include increased agility, flexibility to scale dynamically on demand and reduced time-to-market in rolling out new services.

It offers an edge to operators on the technical front as well. In existing messaging infrastructures, the application software deployed is still linked with the underlying operating system and hardware. Typically, applications are deployed on a one-per-cluster basis. Also, typically, several messaging applications function on different hardware and platforms, such as Solaris Scalable Processor Architecture (SPARC), Intel x86 etc. This requires upgrades to take place at the end-of-life stage of any hardware, coupled with a long procurement cycle, deployment and migration.

Secondly, an underlying challenge is that many messaging applications in use today can be scaled only horizontally. However, the process of upgrading the same is a manual one. This, naturally, requires adequate planning and considerable manpower. By deploying NFV, the network’s functions become wholly software-based, therefore eliminating any dependence on hardware. The software-centric characteristic of virtualized network functions makes it easy to automate deployment, configuration and enable new features like auto-scaling. In fact, network capacity, too becomes flexible and can be added and/or removed from individual functions on demand. One can create a pool of standard x86-based hardware, which can be used for all applications. It is, naturally, cheaper to maintain legacy services like text messaging, while providing a lower entry barrier for new services like chatbots. Automated scaling up and down of resources based on dynamic conditions such as increasing or decreasing traffic result in efficient resource utilization.

However, a challenge is that certain applications cannot be scaled in a linear fashion by augmenting the same with additional hardware. This is where virtualized infrastructure plays a major role. Instead of running a single instance of the application on monolithic high-end hardware platforms, one can run multiple virtual machines to enable multiple instances of the application on a similar high end platform. This implies the hardware platform is utilized more efficiently, in terms of capacity, space and power. For example, for large scale deployments, instead of running 20 small servers, opting for five high-end servers and creating multiple virtual machines on these servers can result in significant savings in terms of space and power.


Is an Operator’s Messaging infrastructure NFV-ready?

Interestingly, as per industry reports, existing messaging applications are NOT NFV-ready! This, in particular, is applicable when the existing software is ported to run on a virtualized infrastructure. To counter this, the NFV architecture has to be clearly “defined” from scratch, in order to meet requirements pertaining to scalability, performance, etc.

Moreover, the overall design ought to be modular with re-usable components that focus on optimizing resource usage. Dynamic scaling up and down, based on network conditions and resource utilization becomes a critical requirement. Multi tenancy (hosting multiple operators or organizations) requirements also become important. Operational requirements such as ease of installation, configuration, upgrades and maintenance are equally important.

Best Practices for Deploying NFV in Messaging Applications:

  • Key messaging functions ought to be separated into virtual network functions. This is aimed to ensure that multiple channels such as SMS, USSD, MMS and chatbots are accommodated. These channels need to the ability to scale independently as well.
  • Various virtual network functions can be grouped together to establish service chains. This is to ensure the infrastructure has omni-channel enterprise messaging capabilities.
  • Defining the individual virtual network functions to comprise of components (VNFCs) capable of handling unique functions such as content filtering, etc. The component based structure of the virtualized network functions ensures that the components can scale independently and also be upgraded independently of other components
  • The virtualized network functions ought to be, modular, scalable and easy to automate deployment and configuration
  • Ensuring that VNF interfaces with management and orchestration-functions are in line with the recommendations of the Network Functions Virtualization Industry Specification Group at European Telecommunications Standards Institute. This, in turn, is aimed at ensuring ease of interoperability with various management and orchestration solutions.


Guruprasad Balasubramanyam

Guruprasad Balasubramanyam

Guruprasad is Senior Architect at Messaging and Broadband solutions unit within Mahindra Comviva. He current focus is on Virtualizing Network Functions and virtualized deployments of products. He has 16+ years of telecom industry experience...