AWS re: Invent 2018: Amazon unveils its new Cloud services
As part of the AWS re: Invent 2018 conference which takes place from November 26 to 30 in Las Vegas, Amazon has unveiled new Cloud, Machine Learning and Big Data services for its Amazon Web Services platform. We take stock of the most significant announcements.
In the Cloud Computing market, Amazon Web Services is the undisputed leader . The firm alone has a 33% share of the public cloud market. Determined to establish its domination, the American company has just unveiled its new Cloud, Machine Learning and Big Data services as part of the AWS re: Invent 2018 conference.
The principal objective ? Reduce costs, accelerate machine learning algorithm training, and increase Cloud Computing performance. Here are the highlights of the conference .
AWS re: Invent 2018: Graviton chips for Data Centers
Amazon Web Services has created its own chips for Cloud Computing . Graviton chips are based on the same ARM architecture as smartphone and tablet processors, and are created by the team that joined the company in 2015 following the acquisition of the startup Annapurna Labs. AWS ‘objective is better software and hardware integration in its Data Centers in order to be able to offer new, cheaper services.
These chips are specifically designed for maximum efficiency in AWS Data Centers, which reduces costs. Users will be able to choose to rent servers equipped with Graviton chips via the new virtual machine instances A1 on the EC2 platform. On certain services, the bill may be reduced by up to 45% compared to servers equipped with AMD or Intel chips.
AWS re: Invent 2018: P3dn instances accelerate Machine Learning
Thanks to the new P3dn instances that will replace the P3 instances, AWS will make it possible to reduce the training time of Machine Learning algorithms to less than an hour. Indeed, P3dn instances allow transferring four times more data at 100 GBps from Amazon S3 or Amazon EFS to the GPUs used for training. Virtual machines will also benefit from this acceleration with the new C5n instances.
In addition, the hardware of P3dn instances is also improved compared to P3 instances. These infrastructures are based on 8 Nvidia Tesla V100 GPUs, 96 vCPU Intel Xeon Scalable, and 1.8TB of NVM-based SSD storage. As a reminder, Amazon has also opened its free Machine Learning training program to all AWS developers.
AWS re: Invent 2018: Elastic Fabric Adapter service for HPC on the Cloud
In order to convince companies running high-performance computing workloads to come to its Cloud rather than using their own supercomputers, AWS unveils the new Elastic Fabric Adapter service. This allows virtual machines to share data via low latency interlinks.
The EFA service is integrated with the Message Passing Interface, which allows HPC applications to scale to tens of thousands of CPU cores without any modification. This service is available in preview on AWS EC2 P3dn and C5n instances and will be offered on more instances in 2019.
AWS re: Invent 2018: new tools to simplify the creation and execution of IoT applications from the Cloud
For the Internet of Things, AWS unveils a new suite of tools to facilitate the development and execution of IoT applications . First, AWS IoT SiteWise is a managed service that collects, organizes and structures the data collected by IoT devices in industrial complexes. Thus, it is possible to analyze the equipment and performance data.
The A WS IoT Events managed service monitors IoT sensors and applications to detect problems such as machine malfunction. The AWS IoT Things Graph service, on the other hand, allows you to link devices such as IoT sensors to services via a drag-and-drop interface. Finally, AWS IoT Greengrass Connectors allows developers to connect to third-party services like ServiceNow or Splunk through common APIs.
AWS re: Invent 2018: Comprehend Medical extracts important data from medical reports thanks to Machine Learning
The new Comprehend Medical service makes it possible to extract important data from medical reports thanks to Machine Learning. It could save healthcare professionals money , and make better treatment decisions.
This Cloud service is based on Machine Learning, and textual analysis in order to decipher the recordings which generally include prescriptions, notes, audio interviews, and exam results . Once the records have been scanned and downloaded, Comprehend Medical selects the relevant information and organizes it.