Have you noticed the recent up-tick in headlines around telco offerings by hyperscalers AWS, Google and Microsoft? Or the multi-cloud telco models, the middleware, supplied by VMware and Red Hat?
Whilst previous generations of wireless connectivity have focussed on voice and data capabilities, 5G is architected to better enable consumer business models. Edge compute (both on-prem / device-edge and provider edge) and related services to support 5G use-cases appears to be the leading driver behind recent announcements. These use-cases will need to be managed by our OSS/BSS for the telco operators and their customers.
In fact AWS and more recently, Google, have announced their own private network offerings.
Meanwhile, top-tier OSS/BSS users are also continuing to adopt cloud-native OSS/BSS initiatives, as described in this Infographic from Analysis Mason / Amdocs. Analysis Mason estimates that over 90% of CSPs in North America, Asia–Pacific and Europe will have their OSS/BSS stacks running on cloud infrastructure by 2022, with well over 60% on hybrid cloud.
However, just how much of the CSP OSS/BSS stack will be on the cloud remains in question. According to TM Forum’s research, most CSPs have deployed less than 5% percent of their operations software in the public cloud.
In today’s article, we take a closer look into cloud offerings for OSS/BSS. The providers we’ll cover are hyperscalers:
The following diagrams come from the Amazon Telco Symposium. The first diagram shows the AWS Telecom Engagement Model (noting the OSS/BSS bubble).
The latter diagram provides some insight into important offerings in AWS’ push into the 5G / telco edge such as Greengrass, SiteWise, Sagemaker and more.
AWS services such as the following have been used as part of home-grown offerings for years:
- Wavelength (low latency), Lambda (serverless) or EC2 – compute services for processing applications/code
- S3, EFS, Glacier, Elastic, Snow Family, etc – data storage for collecting logs, etc
- EKS or ECS – for Kubernetes / Docker container / cluster management
- VPC – for separate environment deployments
- VPN – to tie VPCs to networks / clouds / DCs
- ELB – for load balancing
- ELK – for log management consisting of three open source projects: Elasticsearch, Logstash, and Kibana
- Aurora, RDS, Redshift, DynamoDB, Neptune, KDB, etc – databases
- Cassandra, Kibana, etc – data visualisation
- SageMaker, Augmented AI, Lex, etc – AI / ML tools
- And much more
These have been leveraged by telco architects to build home-grown OSS/BSS tools that leverage commercial and open-source products like Apache’s Kafka, NiFi, Spark, etc.
However, there’s been an increasing trend for OSS/BSS vendors to publish their offerings on the AWS marketplace too, including:
- Zoho / ManageEngine (eg OpManager, Network Config Mgr)
- Domotz Pro
- Lumeta CloudVisibility
- Mphasis InfraGraf (Forecasting & Planning)
- KloudGin (Work Order Mgmt)
- Kx (network performance)
- AND Bosch, ThingsBoard, ThingPark, ThingLogix, etc, etc (if we extend into IoT device management)
- Rakuten Symworld
- EXFO Adaptive Service Assurance (ASA) test solution and advanced analytics reporting module
AWS Marketplace tends to show the solutions that are more standardised / fixed-price in nature (Telecoms section in Marketplace). Many other OSS/BSS vendors such as Netcracker, CSG, Intraway and Camvio don’t appear in the AWS marketplace but have customisable, AWS-ready solutions for clients. These companies have their own sales arms obviously, but also train the AWS global salesforce in their products.
The article provides insight into how Netcracker’s Active Inventory has been implemented over AWS infrastructure.
The following content on AWS in italics also highlights its cloud-based capabilities around generative AI for telco use-cases:
- Choosing the right Foundation Model. Amazon Bedrock is a managed service that provides access to generative AI models from leading AI startups like AI21 Labs, Anthropic, Stability AI, and Amazon’s own Titan models. This enables Telcos to select the perfect model for their required use case. In addition, all models are available through APIs, which makes it easy to build generative AI capabilities into customer and third-party applications. Amazon SageMaker JumpStart offers FMs not available in Amazon Bedrock such as Cohere and LightOn, as well as open source models such as Flan T5, GPT-J and Bloom.
- Saving Time and Money on Foundation Model Training. Amazon Elastic Compute Cloud Trn1 (Amazon EC2) instances powered by AWS Trainium are purpose-built for high-performance deep learning (DL) training of generative AI models. They reduce the time required to train models from months to weeks, or even days, while also lowering costs. This enables Telcos to save up to 50% on training costs versus other EC2 instances.
- Improving Productivity and Reducing Deployment Costs. When deploying generative AI models at scale, most costs are associated with running the models and doing inference. Fortunately, Telco customers can cost-effectively crunch massive amounts of data with the help of Amazon EC2 Inf2 instances powered by AWS Inferentia2. Inf2 instances are optimized for large-scale generative AI applications with models containing hundreds of billions of parameters (and deliver up to 4x higher throughput and up to 10x lower latency than Inf1 instances).
- Building Applications Faster and More Securely. Amazon CodeWhisperer radically improves developer productivity by making coding seamless. The AI coding companion uses a foundation model to generate code suggestions in real-time based on developers’ comments in natural language and prior code in an integrated development environment. It also has built-in security scanning (powered by automated reasoning) for finding and suggesting remediations for hard-to-detect vulnerabilities.
According to Google Cloud’s strategy for the telecom industry, Google Cloud is focusing on three strategic areas to support telecommunications companies:
Helping telecommunications companies monetise 5G as a business services platform, including:
The Global Mobile Edge Cloud (GMEC) strategy, which will deliver a portfolio and marketplace of 5G solutions built jointly with telecommunications companies; an open cloud platform for developing network-centric applications; and a global distributed edge for deploying these solutions
- Anthos for Telecom, which will bring its Anthos cloud application platform to the network edge, allowing telecommunications companies to run their applications wherever it makes the most sense. Anthos for Telecom—based on open-source Kubernetes—will provide an open platform for network-centric applications.
Empowering telecommunications companies to better engage their customers through data-driven experiences by:
Empowering telecommunications companies to transform their customer experiences through data- and AI-driven technologies. Google’s BigQuery platform provides a scalable data analytics solution—with machine learning built-in so telecommunications companies can store, process, and analyze data in real time, and build personalization models on top of this data
- Contact Center AI assists telecommunications companies with customer service. Contact Center AI gives companies 24/7 access to conversational self-service, with seamless hand-offs to human agents for more complex issues. It also empowers human agents with continuous support during their calls by identifying intent and providing real-time, step-by-step assistance
- AI and retail solutions including omni-channel marketing, sales and service, personalisation and recommendations, and virtual-agent presence in stores
Assisting them in improving operational efficiencies across core telecom systems. This allows operators to move OSS, BSS and network functions from their own environments to the Google Cloud
This LightReading report even highlights how Google has been engaged to provide extensive knowledge transfer to some telcos.
This press release from March 2020 announced that Google would partner with Amdocs to support the telecom industry to:
- Deliver Amdocs solutions to Google Cloud: Amdocs will run its digital portfolio on Google Cloud’s Anthos, enabling communications service providers (CSPs) to deploy across hybrid and multi-cloud configurations
- Develop new enterprise-focused 5G edge computing solutions: Amdocs and Google Cloud will create new industry solutions for CSPs to monetize over 5G networks at the edge
- Help CSPs leverage data and analytics to improve services: Amdocs will make its Data Hub and Data Intelligence analytics solutions available on Google Cloud. Amdocs and Google Cloud will also develop a new, comprehensive analytics solution to help CSPs leverage data to improve the reliability of their services and customer experiences.
- Partner on Site Reliability Engineering (SRE) services: The companies will share tools, frameworks, and best practices for SRE and DevOps
On the same day the Google / Amdocs partnership was announced, Netcracker Technology announced it would deploy its entire Digital BSS/OSS and Orchestration stack on Google Cloud. These applications are cloud native, deployed as a set of reusable microservices that run over on-prem or public cloud on top of container platforms such as Google Kubernetes Engine (GKE).
Optiva (formerly Redknee) has also adopted a Google Cloud strategy, using Google Cloud Spanner and the Google Cloud Platform (GCP) to underpin its Charging Engine.
Ericsson has also announced a partnership with Google Cloud, to jointly develop 5G and edge cloud solutions as well as enterprise applications at the edge. Similarly, Nokia has also partnered with Google Cloud to collaborate on RAN / Open RAN, Cloud RAN and edge cloud technology / apps.
This article from CIMI Corp provides some great additional reading about why Google is a credible telco cloud provider.
In June 2022, Google Cloud has even announced a new private networking solution portfolio for private cellular networks. Shortly after, it announced this major activity working in conjunction with Vodafone on an AIOps platform.
Microsoft has also announced an intention to better serve telecom operators at the convergence of cloud and comms networks through its Azure platform.
The diagram below, from this Microsoft blog, shows their coverage map of offerings for CSP customers (Azure for Operators):
The blog also indicates that their offering is built upon:
- Interconnect – 170 points of presence and 20,000 peering connections globally. More than 200 operators are already integrated with the Azure network via their ExpressRoute service
- Edge – offered by the Azure platform, whether at enterprise edge, network edge, network core or cloud
- Network Functions – This is where Microsoft distinguishes itself from AWS and Google. Its ability to offer network, particularly for 5G via RAN and Mobile Packet Core offerings (see more about Microsoft’s Affirmed and Metaswitch acquisitions below)
- Cloud – Incorporates a marketplace with capabilities including OSS/BSS, IoT (via IoT Central), Machine Learning / AI, Azure Cognitive Services (APIs/SDKs that help developers build cognitive intelligence across Decisions, Vision, Speech, Language and Web-search)
“We will continue to partner with existing suppliers, emerging innovators and network equipment partners to share roadmaps and explore expanded opportunities to work together, including in the areas of radio access networks (RAN), next-generation core, virtualized services, orchestration and operations support system/business support system (OSS/BSS) modernization,” states Yousef Khalidi in this Microsoft post.
Acquisition of Metaswitch Networks (a provider of virtualised network software) and Affirmed Networks (a provider that sells virtualised, cloud-native mobile network solutions) shows further evidence of ambitions in the telco / cloud domain.
Like the partnerships described with AWS and Google above, Netcracker has also partnered with Microsoft, offering its Netcracker Digital BSS/OSS and Orchestration applications on Microsoft Azure. This article also describes that, “Netcracker is collaborating with Microsoft to integrate Azure Machine Learning (ML) and AI services with Netcracker’s Advanced Analytics to add intelligent contextual decisioning and recommendations to enable more personalized customer engagements.”
Meanwhile, Amdocs and Microsoft have been working on making ONAP available on the Azure platform. Nokia and Microsoft are partnering on “…cloud, Artificial Intelligence (AI) and Internet of Things (IoT), bringing together Microsoft cloud solutions and Nokia’s expertise in mission-critical networking.”
Microsoft Azure Edge Zones are offered through Azure, with select carriers and operators, or as private customer zones. They bring compute, storage, and service availability closer to the customer / device for low-latency + high-throughput use cases.
The recent announcement of an AT&T and Microsoft alliance (and more recently here where AT&T will run its 5G mobility network over the Azure cloud) as well as deals involving Telefónica (with Aura, its AI-powered digital assistant), SK Telecom (5G-based cloud gaming), Reliance Jio (cloud solutions), NTT (enterprise solution offerings), and Etisalat (future networks) show an increasing presence for Azure within the telco domain.
Netcracker has integrated its solution with Microsoft business applications (eg Office 365, Dynamics 365 and OneDrive), as other OSS/BSS providers undoubtedly have too.
Microsoft also has their own OSS/BSS offering in Dynamics 365 Field Service Management.
VMware Telco Cloud
VMware’s recently announced 5G Telco Cloud Portfolio has been designed to give network operators the platform to accelerate 5G and Edge implementation. Its key differentiator from the examples provided above is it allows operators to run containerised workloads across private, telco, edge and public clouds. This is seen as being an important feature allowing telcos to avoid cloud partner lock-in.
The press release above indicates that, “VMware is evolving its VMware vCloud NFV solution to Telco Cloud Infrastructure, providing CSPs a consistent and unified platform delivering consistent operations for both Virtual Network Functions (VNFs) and Cloud Native Network Functions (CNNFs) across telco networks. Telco Cloud Infrastructure is designed to optimize the delivery of network services with telco centric enhancements, supporting distributed cloud deployments…
Tightly integrated with Telco Cloud Infrastructure, VMware’s Telco Cloud Automation intelligently automates the end-to-end lifecycle management of network functions and services to simplify operations and accelerate service delivery while optimizing resource utilization. Telco Cloud Automation also now supports infrastructure and Containers-as-a-Service (CaaS) management automation to streamline workload placement and deliver optimal infrastructure resource allocation. It also significantly simplifies the 5G and telco edge network expansions through zero-touch-provisioning (ZTP) whenever capacity is required.“
Red Hat’s origins, of course, are in open-source tools such as Red Hat Linux. They’ve now evolved to become a leading provider of open-source solutions for enterprise, delivering Linux, cloud, container and Kubernetes technology. Red Hat was acquired by IBM in 2019.
Red Hat’s telco offerings are built upon the premise that service providers will use more open-source, multi-vendor solutions to underpin their OSS, BSS and networks of the future. Red Hat aims to offer open infrastructure to facilitate service provider initiatives such as NFV, 5G, OpenRAN and Edge Compute. This includes coordination of telco and IT infrastructure, but also the applications and data that are intertwined with them.
Red Hat’s telco proposition is supported by:
- OpenShift – a cloud compute platform as a service (PaaS), built upon Docker containers (containerised applications) that are managed by Kubernetes. This is particularly relevant to support telco cloud models that provide virtualised network functions. It also helps to deliver the edge compute infrastructure that’s becoming synonymous with 5G
- OpenStack – a set of components, mostly deployed as infrastructure as a service (IaaS) that help manage compute, storage and networks. Some of the components are shown in the diagram below sourced from redhat.com
- Ansible Automation Platform – to automate network configuration, fault remediation, security updates, and more
- Marketplace – to assist service providers in finding, buying and deploying certified, container-based software
- Telco Ecosystem Program – that brings together enterprise and community partners to deliver integrated telco solutions. Partners include Affirmed, Altiostar, Atos, Cisco, Ericsson, Amdocs, MYCOM OSI, Zabbix, Metaswitch, Nokia, Juniper and more
- Consulting – offering service resources that include Innovation Labs, training and consulting
- And other solutions such as Ceph Storage, Cloud Suite, Quay, JBoss suite, Integration, Insights, Fuse and more.
CSPs (Communications Service Providers) find themselves in a catch-22 position with cloud providers. Their own OSS/BSS and those of their suppliers have an increasing reliance on cloud provider services and infrastructure. Due to economies of scale, efficiency of delivery, scalability and a long-tail of service offerings (from the cloud providers and their marketplaces), CSPs aren’t able to compete. Complexity of public cloud (security, scalability, performance, interoperability, etc) also make it a quandary for CSPs. It’s already a challenge (commercially and technically) to run the networks they do, but prohibitively difficult to expand coverage further to include public cloud.
Yet, by investing heavily in cloud services, CSPs are funding further growth of said cloud providers, thus making CSPs less competitive, but more reliant, on cloud services. Telco architects are becoming ever more adept at leveraging the benefits of cloud. An example is being able to spin up apps without having to wait for massive infrastructure projects to be completed first, which has been a massive dependency (ie time delay) for many OSS/BSS projects.
In the distant past, CSPs had the killer apps, being voice and WAN data. These services supported the long-tail of business (eg salespeople from every industry in the world would make sales calls via telephony services) and customers were willing to pay a premium for these services.
The long-tail of business is now omni-channel, and the killer apps are content, experiences, data and the apps that support them. Being the killer apps, whoever supplies them also takes the premium and share-of-wallet. AWS, Google and Microsoft are supplying more of today’s killer apps (or the platforms that support them) than CSPs are.
The risk for CSPs is that cloud providers and over the top players will squeeze most of the profits from massive global investments in 5G. This is exacerbated if telco architects get their cloud architectures wrong and OPEX costs spiral out of control. Whether architectures are optimal or not, CSPs will fund much of the cloud infrastructure. But if CSPs don’t leverage cloud provider offerings, the infrastructure will cost even more, take longer to get to market and constrain them to local presence, leaving them at a competitive disadvantage with other CSPs.
If I were a cloud provider, I’d be happy for CSPs to keep providing the local, physical, outside plant networks (however noting recent investments in local CSPs such as Amazon’s $2B stake in Bharti Airtel and Google’s $4.7 billion investment in Jio Platforms* not to mention Google Fiber and sub-sea fibre roll-outs such as this). It’s CAPEX intensive and needs a lot of human interaction to maintain / augment the widely distributed infrastructure. That means a lot is paid on non-effective time (ie the travel-time of techs travelling to site to fix problems, managing resources and/or coordinating repairs with property owners). Not only that, but there tends to be a lot of regulatory overhead managing local infrastructure / services as well as local knowledge / relationships. Federal Governments want to ensure all their constituents have access to communications services at affordable prices. Local Governments / councils have lots of approvals to get certified. All the while, revenue per bit is continuing to drop, so merely shuffling bits around is a business model with declining profitability.
With declining profitability, operational efficiency improvements and cost reduction becomes even more important. OSS/BSS tools are vital for delivering improved productivity. But CSPs are faced with the challenge of transforming from legacy, monolithic OSS/BSS to more modern, nimble solutions. The more modular, flexible OSS/BSS of today and in future roadmaps are virtualised, microservice-based and designed for continuous delivery / DevOps. This is painting CSPs into a cloud-based future.
Like I said, a catch-22 for CSPs!
You could also see the cloud providers as being similar to the OEMs like Cisco, Ericsson, Nokia, et al that have been supplying infrastructure to the telcos for years. The cloud providers are simply providing the infrastructure of a more software-defined network. They’re just another supplier for CSPs to partner closely with for delivery of services.
But another interesting strategy by Google is that its Anthos hybrid cloud platform will run multi-cloud workloads, including workloads on AWS and Microsoft Azure. Gartner predicts that >75% of midsize and large organisations will have adopted a multi-cloud and/or hybrid IT strategies by 2021 to prevent vendor lock-in. VMware (Dell) and Red Hat (IBM) are others creating multi-cloud / hybrid-cloud offerings. This gives the potential for CSPs to develop a near-global presence for virtualised telco functions. But will cloud providers get there before the telcos do?
For those of us supporting or delivering OSS/BSS, our future is in the clouds either way. It’s a rapidly evolving landscape, so watch this space.
* Note: Google is not the only significant investor in Jio:
|Silver Lake Partners||1.43||2.08%|
|Adia UAE Sovereign||0.8||1.16%|
|Saudi Arabia Sovereign||1.6||2.32%|