Microsoft has announced a preview of Azure Edge Zones to help companies and developers to gear up for the onset of 5G. The aim is to lay the groundwork for “delivering consistent Azure services, app platform, and management to the edge with 5G”. The announcement spans Azure Edge Zones with carriers, connecting Azure services directly to 5G networks in the carrier’s datacenter, “to enable developers to build optimized and scalable applications using Azure and directly connected to 5G networks, taking advantage of consistent Azure APIs and tooling available in the public cloud”. Standalone edge zones will appear in selected cities over the next year. In parallel, the vendor is announcing a preview of Azure Private Edge Zones, which are a “private 5G/LTE network combined with Azure Stack Edge on-premises.”

Canonical loads up on managed apps

Canonical has announced a series of “managed apps”, whereby the Unix vendor will take on management of key applications for customers across a range of public clouds as well as bare metal and private cloud. The vendor said it will take on managing databases including MySQL, InfluxDB, PostgreSQL, MongoDB and ElasticSearch, the NFV management and orchestration application, Open Source Mano, and event streaming platform, Kafka. The offering also spans Prometheus, Grafana, and Graylog.

Ververic goes on the Flink

Ververica, the company formerly known as data Artisans, has launched a community version of its eponymous Apache Flink platform. Apache Flink is a distributed streaming data-flow engine, whose founders also started Ververica. The platform is free for commercial and production use, with no limit on the number or size of the managed Apache Flink applications. It supports “all major hosted Kubernetes services” as well as OpenShift, and offers both a web US and a REST API.

AWS says listen to the GAN

If you’ve exhausted your playlists during lockdown, you could kill at least a few minutes with AWS’s DeepComposer, which has now hit general availability. The platform uses Generative Adversarial Networks to build new samples from an existing dataset. In this case, would be composers or AI experts, play a tune on a musical keyboard, and select a pretrained model for a particular genre. DeepComposer then serves up a “new polyphonic composition based on your tune, by pitting  two different neural networks against each other to produce original digital works. Which makes a change from relying on the guitarist and lead singer butting heads to fill the next album.