“One of the areas we are focusing on is generative AI. We believe that the full benefit and impact of any technology can only be realized once it is democratized.”
Kevin Miller, Vice President of Global Data Centers at Amazon Web Services (AWS), emphasized the importance of driving economic development by accelerating the use of cloud services in an interview with Economy Chosun on Sep. 18. Miller joined AWS in 2008 and led the development of Amazon S3, the company’s storage service.
“AWS has been reacting to new opportunities by building large scale data centers for more than 15 years,” he said. “We recognized early on that specialized hardware optimized for cost and energy consumption would be critical for supporting this growth innovation, and one of the most visible ways we are using innovation to improve power efficiency is in our investment in AWS chips.”
Amazon has launched AI chips to catch up with Microsoft (MS) and Google in the generative AI race, and its subsidiary AWS offers generative AI services that can utilize these chips. AWS offers more than 240 cloud-related services powered by machine learning, artificial intelligence, the Internet of Things, virtual reality, and augmented reality.
How does the AWS data center work?
“AWS has the concept of a Region, which is a physical location around the world where we cluster data centers. Each Region is composed of multiple Availability Zones (AZ), each of which is one or more data centers. Each AZ has independent power, cooling, and physical security and is connected via redundant, ultra-low-latency networks. AWS customers focused on high availability can design their applications to run in multiple AZs to achieve even greater fault-tolerance.”
Could you specify how using AZs can help customers achieve greater fault-tolerance?
“If an application is partitioned across AZs, companies are better isolated and protected from issues such as power outages, lightning strikes, fires, tornadoes, earthquakes, and more. AZs are physically separated by a meaningful distance, many kilometers, from any other AZ, although all are within 100 km (60 miles) of each other. We’re pleased that the AWS Region and AZ model has been recognized by Gartner as the recommended approach for running enterprise applications that require high availability.”
How many of these data centers does AWS operate globally?
“Today, AWS delivers services from 102 AZs across 32 geographic regions around the world. In Korea, this includes an AWS Region in Seoul, which we launched in 2016 to meet the high demand by our tens of thousands of active customers in Korea. Within the AWS Seoul Region, we have four AZs that are engineered to be isolated from failures in other AZs.”
What is AWS’s approach to scaling and optimizing its global data center infrastructure?
“We have focused on continually expanding our services to support virtually any cloud workload, and AWS now has more than 240 fully featured services, including compute, storage, databases, networking, analytics, robotics, machine learning (ML) and artificial intelligence (AI), Internet of Things (IoT), mobile, security, hybrid, virtual and augmented reality (VR and AR), media, application development, deployment, and management.”
Tell us about AWS’ strategy in Korea.
“Our services are designed to scale as customer demand grows, and so our worldwide strategy is consistent in being responsive to local needs. In Korea, due to our high customer demand for low-latency applications, we focused on bringing the cloud closer to the edge that can process large amounts of data faster, and drive productivity gains.
Seoul now has several Amazon CloudFront edge locations that give businesses and developers an easy and cost-effective way to distribute content with low latency and high data transfer speeds, leveraged by many of our gaming, consumer electronics, SaaS, and web portals customers.”
You’ve been with AWS for over 15 years - do you have any memories that stand out?
“Until recently I led the Amazon Simple Storage Service (S3), which stores trillions of files and processes more than 100 million requests per second. Amazon S3 is at the core of hundreds of thousands of data lakes, a centralized repository that allows you to store all your structured and unstructured data at any scale. These data lakes are helping customers of all sizes.
In Korea, one of my favorite stories is about local farmers in Geumsan County who are using AWS storage and data services to grow more perilla leaves. Farmers collect information from various IoT devices installed within the smart farm such as perilla leaf crop images, soil conditions, humidity, rainfall, and other crop-related statistics, which enables them to apply ML to make predictions on environmental changes and generate recommendations for farmers.”
How does AWS work to adopt green practices within its data centers?
“To build an environmentally sustainable business for our customers and the world, we are designing data centers that provide the efficient, resilient service our customers expect while minimizing our environmental footprint—and theirs.
AWS’s scale allows us to achieve higher resource utilization and energy efficiency than the typical on-premises data center. International analyst firm 451 Research, part of S&P Global Intelligence, found across several geographic regions that moving on-premises workloads to AWS can lower the workload carbon footprint by nearly 80%, and up to 96% once AWS is powered with 100% renewable energy by 2025. Cloud data centers are five times more energy efficient than the average on-premises data centers in Asia Pacific.”
How does AWS plan to power all its operations with 100% renewable energy?
“Amazon contracts renewable power from utility scale wind and solar projects that add clean energy to the grid. Globally, for the last three years running, Amazon has been the world’s largest corporate buyer of renewable energy, with over 400 projects and over 20 Gigawatts of installed capacity to date. In 2022, we reached 90% renewable energy across our businesses at Amazon.
AWS will be also water positive by 2030, returning more water to communities than it uses in its direct operations. Our water efficiency metric of 0.19 liters of water per kilowatt-hour demonstrates AWS’s leadership in water efficiency among cloud providers.”
In what ways can companies use AWS’ services to promote energy efficiency and sustainability?
“Our customers are leveraging emerging technology such as AI, ML and IoT to innovate for sustainability. For example, Ecube Labs, a Korean waste management startup, has created an AWS-based IoT solar-powered trash can called ‘CleanCUBE’ that compresses waste to increase loading capacity by up to five times and reduces waste collection frequency by up to 80%. Installing over 8,840 green trash cans in 24 countries, Ecube Labs continues to develop waste management products and services to provide smart, sustainable solutions for future-oriented cities.
Additionally, Nano satellite constellations developer and satellite data services provider Nara Space uses Amazon SageMaker’s ML to improve satellite image quality up to three times, help organizations in tracking carbon footprint emissions, prepare for natural disasters, monitor crop yields and vegetation health, efficiently manage forest and marine resources, detecting illegal fishing boats, and more.”
As new technologies such as AI, IoT, and 5G continue to evolve, how do you envision these trends will impact the design and operation of AWS data centers in the future?
“We’ve been reacting to new opportunities through building large-scale data centers for more than 15 years. One of the most visible ways we are using innovation to improve power efficiency is in our investment in AWS chips. In 2022, AWS launched AWS Trainium, a high-performance ML chip designed to reduce the time and cost of training generative AI models - cutting training time for some models from months to hours. This, in turn, means building new models requires less money and power, with potential cost savings of up to 62% and energy-consumption reductions of up to 29%, versus comparable instances.
Similarly, Inferentia chips are purpose-built to run DL models at scale, helping customers meet their sustainability goals when deploying ultra-large models. AWS’s latest Inferentia2 ML chip is up to 54% more energy efficient and can reduce costs by up to 90% versus comparable instances.”
What are some partnerships that AWS is involved in to stay at the forefront of data center technology and innovation?
“In 2015, AWS acquired Annapurna Labs, a specialist microelectronics company, which is responsible for building innovation in silicon and software for AWS customers. A team of specialist engineers, computer scientists, operations and logistics experts are constantly reimagining computer hardware design - from how to optimize processing performance to protecting components in large, complex devices such as servers when they’re being transported and installed.
In Korea, we teamed up SK Telecom, launching the first AWS Wavelength Zone on the telecommunications provider’s 5G network in 2020 and a second one in 2022. This enables developers to build applications that require single-digit millisecond latency connectivity to mobile devices and end users on SK Telecom’s 5G network, allowing customers to build ultra-low latency applications at the edge of 5G networks for smart factories, interactive live streaming, autonomous vehicles, connected hospitals, and augmented and virtual reality-enhanced experiences.”
What are AWS’ future goals for its data centers?
“AWS is committed to driving economic development across the world by continuing to invest in cloud infrastructure. Accelerating the adoption of cloud services in a nation can create a multiplier effect across many areas of the economy, from job creation and improving productivity, to strengthening operational resilience, and enabling greater agility. Our investment in the construction and operation of cloud infrastructure has created direct and indirect jobs locally, and our efforts continue to have a significant positive contribution to GDP.
At AWS, we believe that the full benefit and impact of any technology can only be realized once it is democratized, and one area we are focusing is on generative AI, which has the potential to be one of the most transformational technologies of our lifetime. We’re working to make generative AI available to more than just large enterprises and well-funded tech companies, and ensuring we have the purpose-built hardware to make training and deploying models faster, easier, and less expensive is important to this effort.”
