Insights from the Session「Harness the power of AWS to build game changing applications」 #AA-04 #AWSSummit

Insights from the Session「Harness the power of AWS to build game changing applications」 #AA-04 #AWSSummit

Clock Icon2024.06.26 09:59

Introduction

Hello, this is Hemanth from the Department of Alliance. In this blog, I’ll share my experience from the AWS Summit Tokyo 2024 session on "Harnessing the Power of AWS to Build Game-Changing Applications". It was fascinating to see that many sessions at AWS Summit Tokyo 2024 offered content in both Japanese and English, making the event more accessible and engaging.

Agenda

  • Game-changing” applications running on AWS
  • Key points for decision makers
  • Why customers choose AWS
  • The first step towards innovation

Bringing Unprecedented Ideas to the Real World

During the session, several innovative applications of AWS technology were highlighted. Each showcased the transformative power of AWS in solving complex challenges:

Astrazeneca

AstraZeneca is utilising AWS's cloud capabilities to expedite the process of drug discovery and development. By utilizing high-performance computers, they can run simulations and handle data more effectively, which drastically cuts down on the time and expense involved in introducing new medications to the market.

which is very inspiring to see how cloud technology is shortening the path from compound discovery to patient treatment, potentially saving lives faster than traditional methods.

Wisk

Wisk does intricate simulations for completely autonomous eVTOL (electric vertical takeoff and landing) aircraft using Amazon Web Services. Significant processing power is needed to guarantee efficiency and safety during autonomous flight.

This highlights how AWS’s scalable computing power can handle the demanding needs of next-generation aviation technologies, paving the way for safer and more efficient air travel.

Leonardo.Ai

With the help of AWS, Leonardo.Ai is revolutionizing content production by producing 4.5 million photos daily. This partnership demonstrates how well AWS can handle complex generative AI applications.

The scale at which Leonardo.Ai operates is a testament to AWS's capability to handle massive data and compute requirements, enabling rapid innovation in content creation.

I was amazed how they were involved in solving all the extremely complex problems in world with technology.

Emergence of new system architecture

The session emphasized the shift from monolithic, on-premises architectures to microservices running in the cloud. This transition allows for virtually unlimited resources and more flexible application management. In today's fast-paced tech industry, moving to a microservices architecture offers enhanced scalability, robustness, and the capacity to develop quickly.

key points for decision makers

  • security and reliability
  • performance and cost optimization
  • maintaining and improving innovation and competitiveness

Secure and reliable cloud

AWS provides robust security frameworks to protect data and applications. Services like AWS Nitro System offer built-in security features that enhance data protection and compliance.

Built in security with AWS Nitro system

AWS Nitro System offers cutting-edge cloud innovation, performance improvements, and security features. It guarantees that client information is safe and unavailable to unauthorized individuals.

Trust of AWS Nitro System from Third Party

  • Independent verification confirms the security of the AWS Nitro System design.
  • Emphasis on "no access by operators" ensures that customer data remains private.
  • Reports from NCC Global validate the system's safety, which you can read more about here

Different Storage Service Features

  • Identity and Access Management: AWS ensures no unauthorized third-party access, and IAM policies securely control access to Amazon EFS. IAM’s robust access controls are crucial in maintaining data integrity and preventing unauthorized access, aligning with best security practices.
  • Tokenization: Provides encryption for data in transit and at rest across various storage services. Encryption at both levels ensures comprehensive data security, making it difficult for potential intruders to access sensitive information.
  • Monitoring and auditing: Continuous monitoring and auditing tools like Amazon CloudWatch and AWS Security Hub provide real-time insights and security oversight. Continuous monitoring is essential for early detection of anomalies, enhancing the overall security posture.

Infrastructure that Delivers Performance and Value

History of AWS Graviton Processor

From Graviton1 (released in 2018) to Graviton4 (released in 2023), AWS has consistently improved performance and efficiency, increasing the number of cores and transistors significantly. This evolution of the Graviton processors highlights AWS’s commitment to delivering cutting-edge performance while optimizing costs.

New Information

  • Graviton 4: This is custom processor brought to you by Annapurna Labs like 50% more CPU cores, 2x cache capacity and has 75% higher memory bandwidth.
  • EC2 R8g instance: The 8th generation instances powered by AWS Graviton4, it's quite amazing as it provides 30% more computing power and highest cost performance on EC2. It also ensures scalability for high-load databases with up to 3 times memory capacity. One of the most energy efficient applications requiring large amounts of memory.
  • EC2 P5 Instance: The highest performance GPU powered instances for deep learning and HPC applications. This is equipped with 8 NVIDIA H100 Tensor Core GPUs which has 4x performance AND 40% lower cost than previous generation. Delivers 8PFlops of computing power, 640GB of high-bandwidth memory and 3200 gbps of EFA network bandiwidth. EC2 UltraClusters with up to 20,000 GPUs achieves computational performance of up to to 20EFlops.
  • AWS trainium Trn1, Trn1n instances: This achieves high performance, cost efficient, and cost effective training processing. 50% most effective training training processing compared to similar EC2 instances. for large scale model training adavantage of upto 64,000 trainium accelerators with EC2 Ultracluster. Has network upto 1600 Gbps deluvers high training performance. It enables end to end machine learning development with popular ML frameworks such as PyTorch, TensorFlow, and JAX.
  • AWS Trainium2: This 2nd generation accelerator designed for generative AI and Machine learning. This is 4x faster than AWS Trainium. It also supports building ultra clusters with 100,000 trainium2 chips. This is ideal for training large-scale generative AI models with billions to trillions of parameters. Seamless Integration with industry standard frameworks such as PyTorch, TensorFlow, JAX, etc.
  • Inf2 instances powered by AWS Inferentia2: It has high performance, high power efficiency and lowest cost inference processing. It is featuring up to 12 inferentia2 accelerators and up to 384GB of HBM2e high speed memory. It can have upto 4x the throughput and up to 1/10x the latency compared to Inf1 instances. It is Optimized for deploying models with over 100 billion parameters.

These advancements make AWS an attractive choice for businesses seeking high-performance computing solutions without compromising on cost-efficiency.

Customer Information

  • SAPHANA says that after implementation it had 30% high compute performance.
  • TuftsMedicine noticed that their performance improved by 68% high performance and also an increase 3x increase number of users per instance.
  • Lacework moved to Amazon EKS to improve performance which resulted in 68% reduction in critical alarms, 66% reduction in compute costs and accelerate data collection and gain faster insights.

Another major information that caught my attention is that to turning machine learning to improve and enhance customer experience. These case studies underscore AWS’s effectiveness in delivering tangible performance benefits and cost savings for various applications.

Continuous Innovation to meet changing needs

This includes a collection of wide ranging and fulfilling services and functions. It has upto 750+ instance types for almost any workload. One such is the Amazon Q: Generated AI-powered assistant brought to you by AWS. Capable generative AI assistant that supports various places which Maximize data value, Developing more functional and secure software, Improving contact center customer service and also expand the supply chain visibility.

Human Resource development accelerates inovation

AWS offers digital training using AWS Skill builder which has 600+ on demand courses, practice risk free with over 1000 lab experiences. Group training to gain experience working with your hands-on environment and gain ability to tackle problems and complex challenges with AWS Experts. Get certified to validate your skills and visualize experts within your organization. Finally AWS education program for ideal content for first step for cloud beginners. which is for reskilling within organization and improving the skills of young people.

Investing in training and development is crucial for leveraging AWS’s capabilities fully, ensuring that organizations remain competitive and innovative.

Conclusion

The presentation gave a fascinating summary of how AWS's technology developments are enabling businesses to reach previously unheard-of performance and innovation levels. AWS provides the resources and assistance required to transform innovative concepts into workable solutions, regardless of the industry—healthcare, aviation, or any other. It is highly recommended that you investigate AWS's most recent offers and think about how you could use them to improve your own endeavors and projects.

  • AWS Summit Tokyo 2024: AWS Summit Tokyo 2024 - Information on sessions, keynotes, and more from the event.
  • AWS Graviton Processors: AWS Graviton - Discover the family of Arm-based processors designed to deliver the best price performance for cloud workloads.
  • AWS Nitro System: AWS Nitro System - Explore the high-performance virtualization infrastructure that powers modern EC2 instances.
  • AWS Trainium: AWS Trainium - Understand how AWS Trainium accelerators are optimized for training machine learning models.

この記事をシェアする

facebook logohatena logotwitter logo

© Classmethod, Inc. All rights reserved.