From Cloud Computing To Edge Computing
Today, these services rely on a global infrastructure of sixteen regions. A new region established in France is also planned for the year 2017.
However, the AWS infrastructure is not limited to these sixteen regions, however important they are. Since 2008, AWS has also built its own Content Delivery Network (CDN), named Amazon CloudFront. This allows customers to improve the performance of their applications by distributing static and dynamic content closer to end users. Thus, thanks to CloudFront, NASA broadcast to the world in 2012 the images of the landing of the robot Curiosity on Mars. This event generated more traffic than the Olympics!
Following the announcement in January of two new points of presence in Europe (Munich and Vienna), CloudFront now has seventy worldwide, including three in France: two in Paris and one in Marseille.
Amazon CloudFront, an asset for platform security
The role of these points of presence is not limited to the simple delivery of content, far from it. They also help increase the security of our customers’ platforms by incorporating protection against distributed denial of service attacks (AWS Shield, announced at re Invent 2016) and web attacks (AWS Web). Firewall application).
Thanks to this defense in depth, it is possible to counter the attack as soon as possible, without letting them reach the platforms hosted in the regions.
Run code on CloudFront Points of Presence
Recently, AWS also gave CloudFront DPs the ability to execute code that handles incoming and outgoing traffic. With this new service called Edge, our customers are now able to trigger Lambda functions that will modify HTTP requests to – or from – the origin, ie the web server that hosts the application. Thanks to Lambda’s serverless architecture, our customers do not have to manage any infrastructure: they just need to deploy the code of their functions, which will be automatically broadcast on the CloudFront points of presence.
Edge allows for example to specialize the content based on the properties of the terminal: if it is a mobile, it is useless to serve very large images at high resolution. A small (and therefore lighter) image is enough, which optimizes the user experience by reducing the transfer time, and saves money by serving fewer data.
This feature can also be used to authenticate the user as soon as possible or to conduct A / B testing by showing different content to different groups of users.
Computing: the infrastructure leaves the data center
Is the first step toward running code outside of AWS regions. However, the variety of uses and infrastructure needs of our customers has led us to go even further and to design services that allow them to process data at the source, from the moment they are produced and without the need to transfer them. to the infrastructure hosted in AWS regions. This also allows for data processing in areas where network connectivity is intermittent or non-existent!
AWS Snowball Edge: A piece of AWS cloud at home
Launched in 2015, AWS Snowball is a portable storage device with a capacity of 100 terabytes. Designed to transfer large amounts of data between a customer’s data center and its AWS infrastructure, it is used for example by Digital Globe, one of the world’s leading providers of imaging, data and terrestrial analysis for transfer petabytes of data to the cloud.
At re Invent 2016, AWS announced Snowball Edge, a new version of Snowball now able to run code locally through an embedded Lambda architecture named AWS Greengrass. This means our customers can now have storage and compute capacity on their own site that can be used with the same APIs they use in their AWS infrastructure. A piece of AWS cloud in their data center, in short.
With this computing capacity comparable to that of a server with 16 cores and 64 GB of RAM, our customers will, for example, be able to apply complex processing on their data before sending them to AWS: compression, format conversion, etc. And without overloading their own on-site infrastructure.
Snowball Edge is already used in contexts particularly criticism. For example, Philips Healthcare deploys it in intensive care units, where medical teams use it to store, process and visualize patients’ vital signs. This guarantees continuity of service and care, even in the event of a major computer failure in the hospital. On another note, the US Department of Defense is using Snowball Edge to rapidly deploy storage and computing capabilities in remote areas.Tags: AWS