Modernizing IMS networks on AWS

IP Multimedia Subsystem (IMS) is a platform for delivering IP-based multimedia services such as voice, video, text, and supplementary services (for example, call barring, message waiting, ad hoc conferences, and more) to fixed and mobile users. With 5G, which is a framework to connect these users to networks such as IMS, the core network (5GC) has been revamped from a reference-point model to a service-based architecture (SBA). 5G brings transformational changes to 5GC with a modular and cloud-native approach. Application-heavy platforms with stateful nodal functions that required one-to-one connections and significant interworking between these functions, have been replaced by a network designed around stateless services using a standard HTTP/2 API. Service-based architecture (SBA) offers several benefits such as enabling network slicing, efficient resource utilization, and simplified network functions (NF) registration, discovery, and inter-NF communication using stateless APIs.

IMS, however, is seeing fewer changes. In 3GGP Release 16, SBA support for IMS nodes has been partially added for Voice Over New Radio (VoNR), with the introduction of features such as Proxy Call Session Control Function (P-CSCF) and Home Subscriber Server (HSS) discovery via NRF. In addition, interfaces on the P-CSCF, Interrogating/Serving (I/S)-CSCF, HSS, and application system (AS) have been enhanced to support this 5GC SBA requirement. However, these changes are supplementary, and the IMS network itself does not directly benefit from the cloud-native benefits of 5GC. IMS is, therefore, transitioning to SBA in steps and is expected to continue to be deployed following a reference-point architectures, using Session Initiation Protocol (SIP) as the underlying protocol to provide multimedia service access in 5G.

In this blog post, we will cover IMS and how you can deploy IMS networks on AWS. We will cover existing IMS deployment issues faced by CSPs (communications service providers) and highlight how you can use services such as AWS Outposts, AWS Local Zones, and AWS developer tools to modernize your IMS workloads heading into 5G and beyond.

IMS deployment challenges

IMS networks have traditionally been deployed on a combination of proprietary hardware and software. Led by European Telecommunications Standards Institute (ETSI) and the Network Functions Virtualization (NFV) framework, telco vendors have since decoupled IMS workloads, and moved some CSPs to virtualized deployments on commercial-off-the-shelf (COTS) hardware. Virtualization is often the first step of transitioning telcos to a cloud-native model, and the benefits of adopting IMS NFV have been immediate in terms of reduced vendor hardware lock-in, and lower total cost of ownership (TCO). However, these virtualized IMS networks are not cloud-native. Our CSP customers still report software cycles taking months, requiring complex deployment operations with little to no automation or adoption of DevOps practices.

NFV provides a framework to virtualize telecommunications NFs such as IMS NFs. However, existing IMS virtual network functions (VNFs) often come with interdependencies that complicate deployments, introduce back vendor lock-in, increase management overhead, and lead to stateful architectures that limit scaling. Some CSPs have attempted to remediate these bottlenecks by exploring on-premises containerized deployments and incorporating DevOps practices. However, you will still find CSPs operating IMS on vendor specific hardware, or a complicated mix of on-premises bare-metal, virtualized, and containerized deployments.

IMS evolution and SBA

With the first 5G framework (Release 16) now standardized, CSPs are deploying 5G networks rapidly to capture the market first, as work continues on Release 17 and beyond. As we move towards 5G SBA, there is opportunity to evolve IMS from the reference-point model to a service-based architecture. IMS SBA will enable stateless and open IMS networks where network functions and services will be exposed to other network functions, not only within the IMS network but the 5GC as well. This interconnection has previously required extensive interworking. You will further be able to add new network functions and services with little to no architectural changes. IMS SBA will also provide the added benefit of being lightweight for the network as well as endpoint devices. In addition, adopting HTTP for multimedia and phasing out SIP will provide reduced signaling overhead and simplified authentication and authorization procedures, making way for onboarding microservices and realizing cloud-native deployment strategies.

IMS SBA is imminent; however, traditional IMS nodes need to be virtualized, or cloud-ready, before they can be integrated into the cloud-native 5G network. CSPs looking to modernize their legacy, often vendor-locked IMS networks, can deploy their virtualized or containerized IMS workloads on AWS and become truly cloud-native. AWS provides on-demand delivery of IT resources such as compute, storage, database, network, and more with pay-as-you-go pricing and no long-term commitment requirements. All AWS customers benefit from the high availability and resiliency of our globally available Regions and Availability Zones. You can further benefit from AWS edge infrastructure services such as AWS Outposts and AWS Local Zones that are designed to bring AWS resources closer to you and your end users. In addition, you can incorporate continuous integration, continuous delivery (CI/CD) pipelines using AWS developer tools such as AWS CodeCommit, AWS CodeBuild, AWS CodeDeploy, and AWS CodePipeline for an end-to-end automated solution based on DevOps practices.

Figure 1. IMS on globally available AWS infrastructure.

IMS on AWS Outposts

AWS Outposts is a fully managed service that extends AWS infrastructure, services, APIs, and tools to your on-premises locations. AWS Outposts provide local access to AWS-managed infrastructure such as Amazon Elastic Compute Cloud (Amazon EC2) for compute, Amazon Elastic Block Store (Amazon EBS) and Amazon Simple Storage Service (Amazon S3) for storage, and Amazon Relational Database Service (Amazon RDS) for relational databases, and more. You can use AWS Outposts to build and run applications on-premises using the same AWS hardware infrastructure, APIs, tools, and management controls that you are familiar with for a consistent hybrid experience.

AWS Outposts is an ideal choice if you’re looking to natively integrate your on-premises environment to AWS services while meeting latency, local data processing, and data residency compliance requirements. You can use AWS Outposts to deploy your IMS workloads on compute, database, and storage servers housed at your facilities and enable seamless and low latency integration with 5GC and legacy networks (for example, LTE, 3G, and others). You can further benefit from native capabilities of managed AWS services deployed on AWS Outposts such as automated Amazon RDS database backups, patching, and storage scaling for Amazon RDS, and container orchestration through Amazon Elastic Container Service (Amazon ECS) and Amazon Elastic Kubernetes Service (Amazon EKS). AWS Outposts connect back to a nearby AWS Region through a service link that allows you to securely connect to your Amazon Virtual Private Cloud (Amazon VPC) and access Regional AWS services and features such as AWS Identity and Access Management (IAM), Amazon CloudWatch for logging and monitoring, Amazon EC2 Auto Scaling groups, and many more.

Figure 2. IMS on AWS Outposts.

You can also use AWS Outposts to help migrate your IMS workloads to AWS. IMS deployments often have latency-sensitive system dependencies that make them difficult to migrate. However, with AWS Outposts, you can segment migrations into smaller parts on-premises while maintaining latency requirements until you are ready to migrate.

AWS Outposts is available as standard 42RU racks. However, there are 1RU and 2RU offerings available as well if you are looking to get started or require a smaller footprint. AWS Outposts is designed to grow with your needs, and you can always add infrastructure such as Amazon EC2 servers, Amazon EBS volumes, and Amazon S3 storage servers, as needed.

IMS on AWS Local Zones

AWS Local Zones are a type of AWS infrastructure deployment that places AWS compute, storage, database, and other select services close to large population areas, industrial centers, and IT hubs. With AWS Local Zones, you can easily run applications that need single-digit millisecond latency closer to your end users in a specific geography. AWS Local Zones are suited for use cases such as media and entertainment, real-time gaming, live video streaming, and more.

AWS Local Zones are ideal if you are looking to complement your existing IMS deployments in large population areas and want to increase availability and capacity in these areas. In addition, you can address your location-dependent use cases and data residency compliance requirements using AWS Local Zones. With AWS Local Zones, you can deploy your IMS stack closer to your end users and provide better quality of service without traffic traversing back to your on-premises locations. With the added benefits of SBA mentioned above, the 5GC can be colocated with the IMS network as well, which opens up new use cases of cross-domain low-latency services exposure and consumption.

AWS Local Zones extend your Amazon VPC from an AWS Region to the Local Zone and allow you to connect to the full range of AWS services (for example, Amazon CloudWatch, Amazon DynamoDB, AWS IAM, and many more) available in the parent Region. You connect to the parent Region over a high-bandwidth, dedicated AWS network backbone using the same APIs and tools that you are familiar with for a consistent hybrid experience. AWS Local Zones also support AWS Direct Connect to route traffic from your on-premises environment to the Local Zone over a dedicated, private connection.

Figure 3. IMS on AWS Local Zones.

Moving to CI/CD for IMS deployments

IMS networks often include cross-platform, cross-vendor interdependencies between NFs such as AS and SCF and networks such as charging, billing, and more. These interdependencies increase management overhead, are error-prone, and a bottleneck to agility and scale. As part of your cloud migration efforts, you can establish CI/CD pipelines for your IMS deployments and reduce these dependencies by introducing automation and breaking silos. We will now cover some of the AWS developer tools available to you.

Any automation pipeline starts with a place to store code. This is where AWS CodeCommit comes into play. AWS CodeCommit is a fully-managed source control service that you can use to securely host highly scalable private Git repositories on AWS. AWS CodeCommit encrypts files at rest and in transit, and is integrated with AWS IAM to control access to repos. AWS CodeCommit works with existing Git tools and supports all Git commands you are familiar with. You can even keep using your preferred development environment plugins, CI/CD systems, and graphical clients. With AWS CodeCommit, you can host all your multi-vendor, interdependent code repositories with fine-grain access controls for each, and benefit from built-in high availability and scalability that allows you to focus on tasks that add value to your business.

For unit testing, telco customers typically set up separate on-premises environments that add management overhead and impact TCO. You can use AWS CodeBuild to compile source code, runs tests, and produce software packages that are ready to deploy. AWS CodeBuild provides prepackaged build environments for popular programming languages and build tools such as Apache Maven, Gradle, and many more. You can even customize these build environments to use your existing build tools and provision as many or few environments as needed.

To automate cloud deployments at scale, you can use AWS CodeDeploy, which deploys software on AWS infrastructure such as Amazon EC2 and Amazon ECS, as well as your on-premises servers for a truly hybrid experience. AWS CodeDeploy is fully customizable through lifecycle hooks, and if you have already incorporated CI/CD pipelines, AWS CodeBuild can be used with existing software release processes or continuous delivery toolchains such as AWS CodePipeline, GitHub, Jenkins, and more.

You can further use AWS CodePipeline to fully automate the entire CI/CD pipeline, where you control all steps to fully automate the build, test, and deploy stages of your code. You even have the option to incorporate your existing build and deployment tools into these pipelines. AWS CodePipeline works out of the box with popular tools like Jenkins, BlazeMeter, and Runscope and allows you to deploy containerized applications to Amazon ECS instances powered by Amazon EC2, AWS Elastic Beanstalk, and even AWS OpsWorks Stacks.

Conclusion

In this blog post, we covered how current IMS networks are deployed, the challenges they face, and how they are likely to evolve with 5G. You can migrate your existing IMS implementations to AWS and benefit from the globally available AWS infrastructure and service features. You can further use AWS Outposts and AWS Local Zones to deploy IMS networks closer to your end users and develop CI/CD pipelines using AWS developer tools to become cloud-native and SBA-ready.

AWS provides several resources to help you get started migrating your telco workloads, such as IMS, on AWS. Get started by visiting the landing zone where you will find reference case studies, solution-specific architectures, blog posts, and much more.