AWS to rely on new emerging tech services for growth: CEO Andy Jassy

Amazon Web Services (AWS), the cloud computing division of Jeff Bezos-led Amazon.com Inc., will depend on new services launched on emerging technologies such as artificial intelligence and blockchain for its next spurt of growth and to stay ahead of its rivals, CEO Andy Jassy said.

Speaking to reporters in Las Vegas, Jassy said that the company was focusing on making solutions more simple for its partners and customers and that the new services were a reflection of the demand that the company had seen in the past year. AWS recorded $27 billion revenue run rate and a growth rate of 46% from a year earlier during the third quarter. 

Jassy also announced a range of new services and hinted that some rivals might copy its offerings. But he said he was confident that the rivals wouldn’t be able to chip away at market share as the company offered more functionalities at lower price points. 

AWS competes with the likes of Microsoft, Google, Oracle and IBM in the cloud computing segment.

AWS on analytics

Jassy said that data was becoming increasingly important for enterprises and that the company had identified the need for these enterprises to form so-called data lakes which could be used to run analytics and get better business intelligence.

In order to help companies with data lakes, the CEO previewed a new service for analytics called AWS Lake Formation.

The new service will make it easy to set up a secure data lake in days rather than weeks or months, Jassy said, adding that developers will be able to ingest, catalog, clean, transform and secure their data. 

AWS Lake Formation will make it easier to combine analytic tools like Amazon EMR, Redshift, Athena, Sagemaker and QuickSight around data in the data lake, the CEO said.

The company also launched Amazon QuickSight, a business intelligence service that will offer machine learning (ML) and natural language-powered insights for all users in an organisation. 

“With ML and natural language capabilities, users can easily discover hidden insights, understand the key drivers, and forecast their business without any ML expertise and countless hours in manual analysis,” the company said in a statement.

AWS on blockchain

Jassy announced two new services for blockchain. He said that AWS had adopted a wait-and-watch policy with the technology as it was overhyped in the market.

“Internally, we were trying to solve issues around blockchain and how databases would work under the new technology. Once we arrived at a solution that we thought would work in the market, we decided to make it public,” he said.

The new blockchain services are Amazon Quantum Ledger Database (QLDB) and Amazon Managed Blockchain.

The QLDB will offer a transparent, immutable and cryptographically verifiable ledger for applications that need a central, trusted authority to provide a permanent and complete record of transactions (for example, supply chain, financial, manufacturing, insurance, and HR).

Amazon Managed Blockchain as a service will make it easy to create, manage and scale blockchain networks for multiple parties to transact in a decentralised manner without the need for a centralised, trusted authority.

Jassy said that the managed blockchain service will work on Ethereum and Hyperledger Fabric frameworks.

AWS on machine learning

The CEO said that machine learning was becoming a necessity for enterprises and launched a flurry of new services.

First up was Amazon SageMaker Ground Truth. The new service, which is part of the company’s one-year-old platform SageMaker for training and deploying machine learning models, is expected to help customers build high-quality and accurate training datasets with an effective combination of automatic labeling integrated with human annotation.

Jassy said the new service will help customers scale ML models faster and cut costs by 70%.

Another service previewed was Amazon Forecast, which is based on the same technology used at Amazon.com.

Amazon Forecast uses machine learning to combine time series data with additional variables to build highly-accurate forecasts, the CEO said, adding that the service doesn’t need any machine learning experience to get started.

Speaking about medical uses of ML, Jassy launched Amazon Comprehend Medical. This service uses machine learning to find insights and relationships in unstructured medical text quickly at scale, the CEO explained. He added that it makes advanced medical text analytics accessible to all developers with no upfront costs.

Next was Amazon Elastic Inference. This service is basically an accelerated compute service that allows you to reduce deep learning inference costs by up to 75% when running inference using Amazon SageMaker or Amazon EC2 instances, the CEO explained. 

Talking about how modern Optical Character Recognition (OCR) technologies had limitations, Jassy unveiled the preview of Amazon Textract. This service automatically detects and extracts text and data from scanned documents. 

“Textract goes beyond simple OCR to also identify the contents of fields in forms and information stored in tables. With Textract you can quickly automate document workflows, enabling you to process a million document pages in one hour,” the CEO said.

Jassy next spoke about how AWS wants to make it easy for developers to create personalisations in machine learning and previewed Amazon Personalize.
“Based on the same personalization technology used by Amazon.com to power product recommendations for millions of customers, Amazon Personalize uses insights generated from customer data to deliver an experience that is tailored to their specific needs and preferences,” the company said in a statement.
The CEO also launched three other services in the form of SageMaker Neo, SageMaker RL and a marketplace for machine learning.

Neo is a new capability of Amazon SageMaker that enables machine learning models to train once and run anywhere in the cloud. SageMaker RL enables support for Reinforcement Learning to enable developers and data scientists to build, train and deploy models in a simulated environment to reflect real-world scenarios in areas such as autonomous systems and robotics.

With Amazon SageMaker Neo, machine learning models can be deployed across different hardware platforms irrespective of the framework on which the model was built, the CEO explained. He added that it optimizes the model with a compiler and a runtime to make the model framework-agnostic and hardware-agnostic.

The AWS Marketplace for machine learning allows customers to build machine learning applications faster with new algorithms and model packages, Jassy said. 

Algorithms and model packages in AWS Marketplace can be deployed directly on Amazon SageMaker, he added.

*The writer is in Las Vegas at the invitation of AWS.

---