The Cloud is Key for the Digital AI Telco Transformation: The Big Interview with Dr. Ishwar Parulkar, AWS Telecom & Edge

The invisible network that connects our ever-expanding array of devices—including smartphones, tablets, laptops, smart home devices, and a myriad of Internet of Things (IoT) gadgets—is on the brink of an extensive global AI-powered digital transformation.

Current studies reveal that the average household now has an average of 21 connected devices, highlighting the significant transformation underway within 5G, 4G, and edge networks in the telecommunications (telco) industry. This shift is set to redefine how we connect, communicate, and interact with our surroundings.

However, this worldwide transformation from cloud to edge faces several challenges, primarily due to the demands of AI and other groundbreaking innovations that require extensive computing resources and ultra-low latency rates.

Techopedia sits down with Dr. Ishwar Parulkar, Chief Technologist of Telecom & Edge Cloud at AWS, to delve into the significance of this digital shift, the critical role of AI within the telco sector, and the essential contribution of cloud providers to this evolving landscape.

About Dr. Ishwar Parulkar

Ishwar Parulkar

Dr. Ishwar Parulkar is the Chief Technologist for Telecom and Edge Cloud at Amazon Web Services, where he shapes the technology strategy, develops new cloud services, and leads efforts to advance AWS’s edge cloud offerings and innovative telecom networks and services.

Before his tenure at AWS, Dr. Parulkar was a Distinguished Engineer at Cisco, acting as Chief Architect for business units focused on telecom routing, mobile packet core, small cell, and network orchestration products. He also served as a Distinguished Engineer at Sun Microsystems, leading the design of pioneering data center computing infrastructures, including the industry’s first multi-core processor systems and compute virtualization platforms.

Advertisements

Dr. Parulkar’s career began at Apple, contributing to the Mac desktop-laptop product lines and the Newton PDA technology, which are viewed as precursors to the iPhone.

Key Takeaways

  • The telco industry is undergoing a significant digital transformation, aiming to integrate AI into networks and edge computing to enhance the capabilities of smartphones, laptops, IoT devices, and more.
  • Efforts are being made to adapt edge networks’ computing resources for the efficient processing of heavy AI workloads, minimizing latency.
  • The global rollout of 5G and the ongoing research and development of 6G highlight the crucial roles of AI and cloud technology in the digital transformation of telco operations.
  • The technology and infrastructure underpinning mobile devices are in a phase of modernization, facing various challenges that demand innovation and advancements.
  • Cloud providers can accelerate 5G and 6G development and help public and private organizations transform their operations with 5G, cloud, and edge technologies.

AI on the Edge: From Cloud to Mobile

Q: AI has taken the world by storm overnight and is moving from the cloud to the smartphone and IoT environment. Is the edge and 5G network infrastructure ready to accommodate the needs of this disruptive technology? Are hardware-software changes needed to update the edge and enable it to process AI?

A: The cloud provides cost-effective storage and data management infrastructure and services, and machine learning (ML) models need large amounts of data to train them. Training models are also compute-intensive, and large language models (LLMs) need clusters of 1000s of GPUs, running for months to train them.

The cloud is the most cost-efficient infrastructure to provide that. However, we do see applications where the inference requires very low latency and, therefore, needs to happen at the edge. Also, the cost of shipping data to the cloud is very expensive and in some cases, it is best utilized at the edge for inference.

For this, we would need GPUs and accelerators in edge hardware along with the ML services running at the edge.

The Role of Cloud Providers in the AI Telco Transformation

Q: What is the role of cloud providers in this digital telco AI transformation?

A: The cloud plays a role in multiple aspects of the telco AI transformation. We see an emergence of telco-specific LLMs from operators as well as vendors. The cloud provides a cost-efficient computing infrastructure that is required to train these large models.

Telcos are building or enhancing existing applications in every area of their business, from customer engagement, internal business operations, and network operations to driving new sources of revenue.

 

The cloud provides tools to work with different foundation models to fine-tune them with additional data and build applications in a secure manner, keeping data private.

And then there are ready-made AI applications for undifferentiated tasks that are available to be used in larger systems that telcos are building.

Open RAN, Virtualization & the New Telco Supply Chain

Q: Open RAN and virtualization of telco hardware have opened up the traditional telco supply chain. How big is this new sector and how can they benefit from the AI-cloud-edge-telco infrastructure?

A: Infrastructure Capex is one of the largest expenditures for telcos. By virtualizing telco hardware and running the software network functions in the cloud or on-premises versions of edge cloud, telcos can reduce the total cost of ownership (TCO).

It also results in an agile network that allows for rapid deployment of new network services. We have had successful production deployments of OSS/BSS, IMS, and 5G core on the cloud already. And now, we are seeing the beginnings of Cloud RAN adoption this year.

For example, at MWC, AWS announced that NTT Docomo will use EKS Anywhere to deploy their Open RAN network.

How AWS Brings the Power of the Cloud to the Edge

Q: How is AWS bringing the cloud to the edge, and for what purpose? Can you give us details on how this concept is put into practice and how telcos benefit from this?

A: While most applications can run in the large central AWS data centers, we see the emergence of a class of applications that need to be closer to the users or on customers’ premises.

For example, applications such as immersive AR/VR, gaming, robotics, autonomous vehicles, and others require very low latency of the order of a few milliseconds; or workloads that need to be maintained on-premises for reasons of data privacy, security, regulatory compliance, and other considerations.

For these applications, AWS built an edge cloud continuum that includes:

  • AWS Local Zones, an AWS-managed infrastructure in AWS-owned sites that sit close to metro locations or industrial centers
  • AWS Wavelength Zones, an AWS-managed infrastructure hosted in data centers owned by telcos at their metro sites
  • AWS Outposts, an AWS-managed infrastructure hosted in a customer’s on-prem data centers

We also have Snow — a family of devices that can be deployed in customer on-prem environments but are managed by the customers themselves.

Each of these edge deployments hosts AWS servers and has a consistent API experience so applications can be written once, deployed anywhere as well as migrated, during deployment across the continuum.

Telcos are using these technologies to host network functions at the edge (on Local Zones and Outposts), build and host private networks (on Outposts and Snow devices), offer public cloud in partnership with AWS (Wavelength), and generate new sources of revenue by building and offering new edge solutions and services (on Outposts, Snow devices, and Wavelength).

Accelerating 5G and 6G with the Cloud

Q: How can the cloud accelerate 5G deployment and drive 6G research and development and standardization?

A: 5G was designed with a service-based architecture and modern interfaces to take advantage of virtualization and enable telcos to build a dynamic, agile network. The cloud can accelerate the deployment taking advantage of this.

By running containerized network functions in the cloud (regions as well as the edge continuum), deploying and managing with a CI/CD (continuous integration/continuous deployment) based life cycle management of the functions, telcos can reduce operations cost significantly.

New network services can be configured and deployed very rapidly using APIs, including creating dynamic network slices, which is one of the end goals of the 5G architecture.

We expect 6G to take this to the next level. While 5G was built assuming a virtualized network implementation, 6G standards will be defined assuming a cloud-native network implementation.

What that means is that the basic definition of 6G will incorporate elements that enable horizontal scaling, elasticity, on-demand provisioning, use of cloud-native services to build network functions, autonomous network operations, and more such cloud design and operation concepts. 6G will also drive a deeper integration of AI technology into all aspects of its architecture and implementation.

Innovation on the Cloud-Edge

Q: What are the most innovative AI-cloud-edge-telco APIs and apps you have seen developed?

A: Each of these technologies — AI and the edge — is evolving rapidly, independently and there are several examples of APIs and apps that either use the edge or ones that use AI technology. And we are seeing some emerging that combine AI and the edge cloud.

While there are several telco APIs being developed, we are seeing early adoption of APIs:

  1. Requesting stable latency (reduced jitter) or throughput for specified application data flows between application clients and servers.
  2. Checking the last time that the SIM card associated with a mobile number has changed.
  3. Performing seamless authentication of a mobile device with the mobile network by requesting a check of the number.
  4. Checking connectivity status for users´ equipment, such as whether it is roaming.
  5. Checking the location of a device.

The most innovative applications we see are in the areas of anti-fraud detection, media streaming, immersive AR/VR experiences, gaming, and autonomous vehicles.

One great example, which won the GSMA Foundry Excellence Award for Digital Transformation and combines edge, AI, and APIs, is Unmanned Life, which offers an orchestration platform for autonomous drones for delivery, surveillance, and data collection. As network performance is critical for remote drone control, Unmanned Life uses Quality on Demand API to request bandwidth and latency guarantees on demand to ensure the seamless operation of their drones.

Challenges That Need to Be Conquered

Q: What are the key challenges that this transformation presents, and why should leaders be looking into them?

A: The adoption of AI and Network APIs is part of a larger shift we refer to as “telco to tech-co,” and the pursuit of the industry to better monetize networks.

We see a few key challenges for telcos: first, telcos need to either develop and hire technical expertise or engage partners.

Telcos such as Telia are investing in upskilling and training their staff, which equips them with the ability to develop their own machine-learning solutions for their customers.

Second, telcos need to become data-centric.

AI is only as good as the data it leverages and how that data is managed. We’re working with telcos to modernize the data platform, manage data using cloud infrastructure, and then layering data analytics, artificial intelligence, and machine learning technologies on top of that data to gain insights, build new offerings, and improve operations.

And finally, telcos need to adopt a new mindset in their operational model, moving from a traditional hardware, static configuration-based, manual operator-heavy approach to a software-driven, dynamic, AI automation-driven approach. This challenge trickles all the way from technology and processes to people.

Yet, most important of all is culture and investment in these shifts from executive leadership. Without that commitment, telcos will not be able to realize the benefits of this new paradigm.

The Future of the New Telco Industry

Q: What will the AI-cloud-edge-telco industry look like in five years?

Communication service providers will continue to face pressure to reduce costs and drive new growth. To address this, we only see the importance of technologies such as generative AI, network APIs, and cloud growth.

In the near term, we see maturation and scale in adoption and even SaaS models for BSS and OSS — many of which are already in production, followed by a similar pattern in IMS and 5G Core.

The last frontier for the cloud in running mobile network functions has been RAN, but this year, we are seeing the first cases of adoption and experimentation.

In the next 5 years, we will see the industry go beyond mobile networks to fixed network functions like BNG (broadband network gateway) and CMTS (cable modem termination systems) and others being run on the cloud.

We see AI play a central and transformational role in the next five years, permeating all aspects of a telco from customer engagement, internal business operations, network operations, and creating new sources of revenue. We will see telco-specific LLMs being built by operators as well as network vendors.

The vision of the network-as-a-service or network as an API-driven platform accessible to developers will come to fruition in the next 5 years as operators start offering APIs and developers unleash their creativity with those APIs to build new applications.

The next 5 years will be critical to the industry, with the early adopters of these powerful, emerging technologies gaining an advantage over others.

Advertisements

Related Reading

Related Terms

Advertisements
Ray Fernandez
Senior Technology Journalist

Ray is an independent journalist with 15 years of experience, focusing on the intersection of technology with various aspects of life and society. He joined Techopedia in 2023 after publishing in numerous media, including Microsoft, TechRepublic, Moonlock, Hackermoon, VentureBeat, Entrepreneur, and ServerWatch. He holds a degree in Journalism from Oxford Distance Learning, and two specializations from FUNIBER in Environmental Science and Oceanography. When Ray is not working, you can find him making music, playing sports, and traveling with his wife and three kids.