News on the Microsoft Azure platform

By now, I think we will have a permanent place on our blog for publications that deal with what's new on the Azure platform. In this summary, I have selected the most important ones that impact the projects being implemented at Hosters.

Not coincidentally, I start this post with solutions related to artificial intelligence, or Azure Machine Learning, which provides tools for machine learning professionals and IT administrators with new features that help streamline and simplify project development.

• Azure Machine Learning registries will help promote, share, and discover machine learning artifacts such as models and environments across multiple organizational workspaces. This will enable customers to track the provenance of models and data across workspaces in different Azure subscriptions, facilitating operations between teams.
• Azure Container for PyTorch and Azure Data Science Virtual Machines for PyTorch, are select environments and custom images that will bundle innovative technologies used by Microsoft to configure, develop, accelerate ,and support optimized training for PyTorch models.

The Responsible AI dashboard in Azure Machine Learning enables customers to quickly implement Responsible AI by debugging machine learning models and making informed, data-driven decisions. Additional updates include:

• Automated machine learning features, including natural language processing and image task support. Automatic code generation is generally available for all automatic machine-learning tasks on tabular, image, and text data.
• The Python SDK v2 simplifies the management of the entire machine learning lifecycle, from training single tasks to datasets and deploying models using Python functions.

The Apache Spark platform in the Azure Synapse Analytics service enables machine learning with big data, providing the ability to derive valuable insights from large volumes of data with structure, unstructured data, and fast-moving data.

Hugging Face will host BLOOM, the world's largest open multilingual language model with 176 billion parameters, on AzureML Managed Endpoints for the community to test performance.

Artificial intelligence is a huge dataset, and the database area is constantly evolving, so recent news is also worth mentioning.

Azure Cosmos DB for PostgreSQL database provides everything developers need to know about PostgreSQL and an advanced extension to Citus for Microsoft's fast and scalable database for developing cloud-native applications. Using the same familiar database service, developers can now create applications with relational and non-relational (NoSQL) data. Additionally, there is support for 16MB file sizes for MongoDB databases.

Azure Database for PostgreSQL - Flexible Server, a fully managed and intelligent PostgreSQL database service, now supports data encryption with client-managed keys. Client-managed keys are stored in the Azure Key Vault service, a centralized critical management service that provides additional security, high availability, and disaster recovery capabilities. The datastore size for the Azure Database for PostgreSQL service will be increased to 32TB. This long-awaited feature best suits workloads running extensive databases or migrating to a flexible server.

Azure Database for MySQL, a fully managed and scalable MySQL database service, can now scale I/O on demand without predefining the number of I/O operations per second.

Another new feature worth mentioning is the Azure Data Explorer service. This fully managed, high-performance big data analytics platform facilitates the analysis of large volumes of data in near real-time. This data management service offers fast data acquisition with low latency and linear scaling, supporting up to 200 MB of data per second per node. The Azure Data Explorer service facilitates remote monitoring of production IoT devices, vehicles, and systems, constantly switching between operations. The toolkit provides customers with a complete data acquisition, query, visualization, and management solution. Data acquisition is the process by which data is added to a table and made available for query execution in the Azure Data Explorer service.

The Azure Data Explorer service provides service from a variety of data sources:

• Data stored within Amazon (S3)
• Azure Synapse Link from Azure Cosmos DB service
• Metrics, logs from open telemetry
• Output from Azure Stream Analytics service

These add-ons make it easy to feed data into the Azure Data Explorer service for near real-time analytics and provide customers with a complete solution for data acquisition, querying, visualization, and management.

SQL Server with support for the Azure Arc service has also gained several new features that enable customers to benefit from a cloud-like environment, including:

• A logging environment that integrates with the Azure Active Directory service (Azure AD)
• Improved security and monitoring with Microsoft Defender enables customers to assess and secure SQL Server resources in multi-cloud hybrid environments.

The new Azure Disk Storage service offering is ideal for applications that require high levels of performance without adding storage capacity. Premium Layer SSD v2 provides the most flexible and scalable general-purpose block storage in the cloud, enabling customers to meet the ever-increasing demands of production workloads such as SQL Server, Oracle, MariaDB, SAP, Cassandra, MongoDB, big data, analytics, gaming, virtual machines (VMs) or stateful containers. With SSDs in the Premium v2 tier, Azure platform customers can now have up to 64 tebibytes (TiB) of storage capacity, 80,000 input/output operations per second (IOPS), and a throughput of 1,200 megabits per second per drive. In addition, customers can now also store incremental snapshots for SSDs at the Premium v2 tier in SSD storage at the Standard tier.

On the topic of service management and monitoring, there are also several new features:

The Azure Automanage service saves time by automating the configuration and management of servers throughout the customer lifecycle at scale, whether on the Azure platform or in hybrid environments supported by the Azure Arc service. The Azure Automanage service provides a set of features for Windows Server virtual machines on the Azure platform that simplify management and reduce downtime of workloads to increase efficiency and cost savings further. One of these features is Hotpatch, which enables patches to be installed without restarting each installation, reducing downtime for Windows Server VMs on the Azure platform.

The new Azure Monitor capabilities are designed to help customers further modernize their environments and optimize costs. These updates include:

• Predictive auto-scaling for Virtual Machine Scale Sets is now generally available. Users can use machine learning to manage and scale the resources of Azure Virtual Machine Scale Sets, predicting the overall CPU load and ensuring that scaling occurs promptly to meet demand.

Azure Monitor Agent (AMA) provides a secure, cost-effective, simplified way to collect telemetry data from your resources. It replaces the older Log Analytics agents (MMA and OMS agents).

Meanwhile, new features for Azure Networking are now generally available and include:

Azure Domain Name System (DNS) private resolver is a cloud native, highly available, and DevOps-friendly service. It provides a simple, low-maintenance, reliable, and secure DNS service for resolving and conditionally forwarding DNS queries from the virtual network, on-premises, and to other target DNS servers without the need to create and manage a custom DNS solution. This will enable customers' DNS infrastructure to run securely, privately, and seamlessly across on-premises networks and enable key hybrid network scenarios.

Azure public multi-access edge compute (MEC) enables enterprises and developers to deliver innovative, high-performance, low-latency applications using operators' public 5G networks. AT&T in Atlanta and Dallas offers Azure public MEC. This offers customers a unique opportunity to analyze data closer to the point of capture for proactive actions and decisions.

Customer Managed Keys (CMK) - provides customers with maximum control over their encrypted vSAN data on the Azure VMware Solution. With this feature, customers use Azure Key Vault to generate customer-managed keys and to centralize and streamline the critical management process, including:

• Revoking access anytime and protecting customer data from within their private cloud.
• Azure Key Vault encryption key rotation as required by customer compliance policies.
• Full integration with Azure Monitor and Azure Log Analytics for audit logs and alerts.

source:microsoft

Summary

Microsoft Azure offers various cloud services that are constantly evolving and adapting to user needs. Azure's catalog of services is continuously being updated and developed, so it's worth keeping up to date with news and updates to realize the platform's full potential. As a Microsoft partner and Azure reseller, we try to choose the best possible solution for our customers' projects, which is why we are constantly expanding our competencies and closely monitoring the platform's development.