This article continues our review of highlights of Azure updates from MS Build 2019.
Video Indexer - multiple facial recognition enhancements have been added
Video Indexer is an Azure service that can be used to extract information from videos. Imagine a scenario where you have to detect the non-celebrity people relevant to a customer’s specific organization. You can use Azure Video Indexer for this job. It has two prominent insights, one is a face detection feature which has the ability to identify 1 million celebrity faces, and the other is a recognition feature which trains the account level of custom person models to recognize the non-celebrity people.
For the latest update in 2019, Azure has made several multiple facial enhancements to Video Indexer which makes the person model training and management faster and easier.
- It now supports 50 person models and each model supports 1 million different people. You can have various use-cases which enable separate and multiple person models for each account.
- Azure allows us to manage face locations through central locations. A “People” tab allows users to view centrally, make customization to existing models, and enables us to index the videos and tag faces in videos.
- Enhanced video indexer has the ability to generate trained Person models from already existing uploaded images of a person. It also allows us to add, delete, create Person models, along with a review, renaming and adding new faces to existing models.
Generating Person Model by using existing uploaded Images
In the last five years, Azure Kubernetes Service has transformed cloud and distributed services. This year with the latest updates, AKS has taken a leap to offer serverless Kubernetes, which gives an opportunity to use AKS virtual nodes using cloud-based serverless containers.
- Now AKS provides fast and serverless containers, running on IaaS infrastructure with no virtual machines to manage, patch and update.
- Kubernetes-based Event-driven Autoscaling has been introduced using FaaS, enabling awareness of event sources and provisions which ultimately helps in matching the resources, now taken at the next level.
Diagram of Kuberbetes cluster
- Azure Dev spaces and Azure Pipelines are working on accelerated development for the developer by minimalizing the dev machine setup. This enables the end-to-end testing in dev environment without mocking their application code.
- AKS has evolved to add security features, enabling policies applied to resources, clusters. These establish the guard rails of behavior for developers to work without permission which enables them to work with DevOps practices. Authenticated IP, which is under preview, restricts the Kubernetes control plane to be accessed by restricted IP.
Azure Kubernetes Service (AKS) virtual node is now available
Azure Kubernetes Service cluster allows you to deploy workloads using virtual nodes. This reduces the use of virtual machines. With the use of virtual nodes, we can have a provision of pods in no time by creating Azure Container instances, with the billing done as per pay per second for their execution time. With this container focused environment in AKS, there is no additional requirement of managing resources.
Azure Stream Analytics support for Visual Studio Code is in preview
Azure Steam Analytics is serverless, fully managed event processing engine that allows analyzing data streams from applications and IoT.
This is used for real-time data analysis for Visual Studio used on Mac, Linux or Windows OS. It is a light-weight, cross-platform editor which enables you to develop, manage and test stream the analytics data.
The analytics data is visualized using Power BI dashboard, while the application is running to analyze the data, Power BI dashboard periodically updates the new arrival data.
Data analysis using Powered BI Dashboard in Visual Studio
Autoscale for Azure HDInsight is now in preview
Azure HDInsight is a cost-effective, enterprise-grade service provided to run popular open source frameworks like Apache, Hadoop, Spark, etc. This tool gets up to run all your development environments on a big data cluster. You can scale them up or down based on your needs.
The latest version allows you to auto scale the HDInsight clusters based on load or according to pre-defined schedules. Load-Based auto scale is intelligent enough to scale HDInsight cluster according to workload-specific metric gathered from CPU; memory usage. Schedule-based auto scale allows setting the custom schedule for scaling. You can set the number of nodes at a given time.
Azure SQL Database managed instance SKU recommendation for Data Migration Assistant is now available
Having the best Azure database target and SKU for databases could be complicated with the migration of the database to the cloud. Data Migration Assistant gives you the power to migrate your database by giving SKU recommendations in user-friendly output.
With the latest release of 2019, SQL server assessments are available to migrate from legacy SQL Database to Azure SQL database managed instances. You can have new Azure SQL database SKU recommendations for cloud database service tier and sizing configurations based on existing workload. This feature recommends the pricing tier, compute level and max data size along with the cost per month. This service is not charged and is free.
Azure VMware Solutions is now generally available
Azure provides VMware Solution which enables the customers to work on VMware environment. With this service, you can deploy the VM on dedicated, single tenant, bare metal infrastructure.
- Managing services, backend infrastructure, networking services and platforms for operations are all provided by Azure.
- VMware by CloudSimple is available in the US East and US West region. This brings the pleasure of working on VM on Cloud environment. Soon it will be delivered to all.
Azure QuickStart Center enables new customers to build cloud projects with confidence
It’s good news for beginners and learners who want to dive in to Azure products. From the 2019 release, Azure QuickStart center enables you to learn with a self-guided approach to start your project on the cloud environment. Quick Start service will be seen in All services or by searching in the search bar.
Using QuickStart you can learn in 3 different ways:
- Setting up your cloud environment using best practices.
- Quickly launching your project and learning the Azure tools for creating the project.
- Take a course online to experience the Azure environment and tools.
Azure Maps Mobility Services is in preview
Azure 2019 updates provide the service of Azure Maps for the integration of real-time public transit networks to work on enterprise level applications. Public transit allows a user to analyze, manage, and optimize public movement through urban and rural areas.
It manages this information using secure geospatial data with built-in location intelligence. It can be used with real-time applications working in IoT and AI solutions. We can enable vehicle routing and traffic simulation using Maps Mobility. Also, it helps in asset tracking services to POI details and satellite image analyzes. It can work with any open source framework.
A real-time example of Maps Mobility
The Cognitive Search feature of Azure Search is now available
With the Cognitive Search Feature announcement in 2018, the ability of developers to analyze the big data and real-time data has deepened. They can have Azure search for facial recognition, key phrase extraction and sentiment analysis.
- This search enables support of complex data including nested data and hierarchical data which are not broken into tabular data.
- Knowledge store capability within cognitive search allows AI-generated annotations in a table or JSON format which uses the non-search user cases like PowerBI dashboard, machine learning models, organized data repository and bot.
- Form Recognizer allows extracting the exact keys and phrases, from tables and documents. This helps to make the decision faster.
Azure SignalR Service support for ASP.NET is now available
Azure SingalR Service helps developer to build an application using real-time environment and features. We can have Azure SingnalR setup with ASP .Net core Identity with serverless Real-time application with Authentication.
Now with the latest release, SingnalR is available to applications for scaling at real-time.
Cross availability zones support for Azure Service Fabric clusters is now available
Azure Service Fabric is a distributed system platform, enabling packing, deployment and managing scalable containers with reliable microservices.
With the Service Fabric cluster available to all, the zones support gives additional power by increasing reliability for the teams and developers working across zones.
Azure App Service (Linux) support for Python is now available
Azure App service offers new built in support for versions of Python, Node.js, PHP on free tiers with the application service on Linux.
- The new build system can deploy the dependencies need for Linux applications as running custom scripts.
- Free Tier is made available for developers of Linux to try out the cloud-development. Developers can even try App services on Linux with selecting free SKU.
- Java 11 is available for Linux and Windows app services. Microsoft is supporting java enabled app services until 2026, which will help developers to play with API and adopt new versions and new changes in newer versions of java.
Azure Container Registry content trust is now available
Both source and integrity must be verified for the security of any distributed system. The consumer must be able to verify the data at the publisher level and published level. Content trust allows signing the images which are published in the Registry and can configure the clients to pull only signed images. When an image is pulled, again the Docker client verifies the integrity of the image. The content trust uses a cryptographic signing key which is specific to the repository they are associated with.
With Container Registry now available to users, several types of signing keys can be used to manage the trust between Docker clients and your registry.