Kategorie: Microsoft Azure

Roundup of AI breakthroughs by Microsoft and NVIDIA

In terms of AI developments, 2022 proved to be a banner year. AI production and usage increased and both deep learning and machine learning models steadily conquered increasingly complex problems. But AI evolution goes beyond impressive production and adoption levels to include breakthroughs in new computing methods, applications, and mega-scale supercomputers. Two of the leaders in this space are Microsoft Azure and NVIDIA and here are some of their most noteworthy achievements this year.

Massive cloud AI supercomputer for mega-sized AI models

The two computing giants are collaborating to provide one of the most powerful AI supercomputing platforms in the world. Powered by Microsoft Azure’s advanced supercomputing infrastructure combined with NVIDIA GPUs, networking, and the full stack of AI software, this cloud supercomputer will help enterprises train, deploy, and scale AI, including large, state-of-the-art models.

Azure is the first public cloud to incorporate NVIDIA’s advanced AI stack, adding tens of thousands of NVIDIA A100 and H100 GPUs, NVIDIA Quantum-2 400Gb/s InfiniBand networking, and the NVIDIA AI enterprise software suite to its platform.

This effort isn’t limited to spinning up better outputs in lean times—there’s also the need to advance knowledge and discoveries to benefit all of mankind’s endeavors and needs. All told, it’s an impressive AI system that performs at scales rarely imagined to be possible.

A quantum leap for AI-fueled scientific discoveries

AI can do far more than traditional data mining and advanced analytics. AI is dramatically changing scientific methods, handling complexities that a human mind cannot comprehend.

At Supercomputing 2022 (SC22), NVIDIA announced broad adoption of its next-generation H100 Tensor Core GPUs and Quantum-2 InfiniBand, including new offerings on Microsoft Azure cloud and over 50 new partner systems for accelerating scientific discovery.

H100 and Quantum-2 are part of NVIDIA’s high-performance computing (HPC) platform—a full technology stack with CPUs, GPUs, DPUs, systems, networking, and a broad range of AI and HPC software—that provides researchers the ability to efficiently accelerate their work on powerful systems, on-premises or in the cloud.

Microsoft Azure will be the first to offer NVIDIA Quantum-2 for HPC Workloads, providing a world-class supercomputing cloud infrastructure that allows researchers and scientists using Azure to achieve their life’s work.

Innovation is the lifeblood of all industries and this Microsoft Azure and NVIDIA breakthrough puts it at the fingertips of all users.

AI brings medical imagery diagnostics into sharper focus

A powerful collaboration between Microsoft Azure, NVIDIA, and the Nuance Precision Imaging Network puts AI-based medical image diagnostic tools directly into the hands of radiologists and other clinicians. This enables the capture of economies at scale, meaning patient care improves while costs drop.

A powerful collaboration between Microsoft Azure, NVIDIA, and the Nuance Precision Imaging Network puts AI-based medical image diagnostic tools directly into the hands of radiologists and other clinicians. This enables the capture of economies at scale, meaning patient care improves while costs drop. 

Powered by Microsoft Azure, the Nuance Precision Imaging Network provides access to an entire ecosystem of AI-powered tools and insights within clinical workflows to more than 12,000 healthcare facilities and the 80 percent of U.S. radiologists who use Nuance's PowerScribe radiology reporting and PowerShare image sharing solutions1. 

Mass General Brigham will be among the first to accelerate end-to-end AI model development and deployment in clinical workflows on the Nuance Precision Imaging Network.

Wrapping up the roundup

Azure collaborates with Hazy Research and NVIDIA to achieve unmatched MLPerf results. Azure-Hazy Research is the only submitter that reached below the 2-minute mark with BERT on 16 virtual machines.
Azure scales 530B Parameter GPT-3 Model with NVIDIA NeMo Megatron. Combining NVIDIA NeMo Megatron with Azure AI infrastructure offers a powerful platform that anyone can spin up in minutes without having to incur the costs and burden of managing their own on-premises infrastructure.
Azure and NVIDIA scored Top500 and Green500 ranking for AI purpose-built infrastructure
HPCwire bestowed Microsoft Azure and NVIDIA top awards in Best Use of HPC in Financial Services Editors’ Choice, Best AI Product or Technology—Readers’ Choice, and Best Use of High-Performance Data Analytics & Artificial Intelligence.
At Supercomputing 2022, Nidhi Chappell, General Manager, Microsoft Azure HPC + AI, and Ian Buck, VP/General Manager Hyperscale and HPC Computing, NVIDIA discussed game changers in AI and cloud infrastructure. And NVIDIA founder and CEO Jensen Huang shared the latest news, innovations, and technologies in NVIDIA’s SC22 Special Address.

Stay tuned for our 2023 advancements

Keep up to date with Microsoft and NVIDIA. Visit our HPCwire Solution Channel.

 

 

Sources: NVIDIA Teams With Microsoft to Build Massive Cloud AI Computer, NVIDIA news, November 16, 2022.
 
Quelle: Azure

Scale Azure Firewall SNAT ports with NAT Gateway for large workloads

This post was co-authored by Suren Jamiyanaa, Product Manager II, Azure Networking.

As large organizations across all industries expand their cloud business and operations, one core criteria for their cloud infrastructure is to make connections over the internet at scale. However, a common outbound connectivity issue encountered when handling large-scale outbound traffic is source network address translation (SNAT) port exhaustion. Each time a new connection to the same destination endpoint is made over the internet, a new SNAT port is used. SNAT port exhaustion occurs when all available SNAT ports run out. Environments that often require making many connections to the same destination, such as accessing a database hosted in a service provider’s data center, are susceptible to SNAT port exhaustion. When it comes to connecting outbound to the internet, customers need to not only consider potential risks such as SNAT port exhaustion but also how to provide security for their outbound traffic.

Azure Firewall is an intelligent security service that protects cloud infrastructures against new and emerging attacks by filtering network traffic. All outbound internet traffic using Azure Firewall is inspected, secured, and undergoes SNAT to conceal the original client IP address. To bolster outbound connectivity, Azure Firewall can be scaled out by associating multiple public IPs to Azure Firewall. Some large-scale environments may require manually associating up to hundreds of public IPs to Firewall in order to meet the demand of large-scale workloads, which can be a challenge to manage long-term. Partner destinations also commonly have a limit on the number of IPs that can be whitelisted at their destination sites, which can create challenges when Firewall outbound connectivity needs to be scaled out with many public IPs. Without scaling this outbound connectivity, customers are more susceptible to outbound connectivity failures due to SNAT port exhaustion.

This is where network address translation (NAT) gateway comes in. NAT gateway can be easily deployed to an Azure Firewall subnet to automatically scale connections and filter traffic through the firewall before connecting to the internet. NAT gateway not only provides a larger SNAT port inventory with fewer public IPs but NAT gateway’s unique method of SNAT port allocation is specifically designed to handle dynamic and large-scale workloads. NAT gateway’s dynamic allocation and randomized selection of SNAT ports significantly reduce the risk of SNAT port exhaustion while also keeping overhead management of public IPs at a minimum.

In this blog, we’ll explore the benefits of using NAT Gateway with Azure Firewall as well as how to integrate both into your architecture to ensure you have the best setup for meeting your security and scalability needs for outbound connectivity.

Benefits of using NAT Gateway with Azure Firewall

One of the greatest benefits of integrating NAT gateway into your Firewall architecture is the scalability that it provides for outbound connectivity. SNAT ports are a key component to making new connections over the internet and distinguishing different connections from one another coming from the same source endpoint. NAT gateway provides 64,512 SNAT ports per public IP and can scale out to use 16 public IP addresses. This means, when fully scaled out with 16 public IP addresses, NAT gateway provides over 1 million SNAT ports. Azure Firewall, on the other hand, supports 2,496 SNAT ports per public IP per virtual machine instance within a virtual machine scale set (minimum of 2 instances). This means that to achieve the same volume of SNAT port inventory as NAT gateway when fully scaled out, Firewall may require up to 200 public IPs. Not only does NAT gateway offer more SNAT ports with fewer public IPs, but these SNAT ports are allocated on demand to any virtual machine in a subnet. On-demand SNAT port allocation is key to how NAT gateway significantly reduces the risk of common outbound connectivity issues like SNAT port exhaustion.

NAT gateway also provides 50 Gbps of data throughput for outbound traffic that can be used in line with a standard SKU Azure Firewall, which provides 30 Gbps of data throughput. Premium SKU Azure Firewall provides 100 Gbps of data throughput.

With NAT gateway you also ensure that your outbound traffic is entirely secure since no inbound traffic can get through NAT gateway. All inbound traffic is subject to security rules enabled on the Azure Firewall before it can reach any private resources within your cloud infrastructure.

To learn more about the other benefits that NAT gateway offers in Azure Firewall architectures, see NAT gateway integration with Azure Firewall.

How to get the most out of using NAT Gateway with Azure Firewall

Let’s take a look at how to set up NAT gateway with Azure Firewall and how connectivity to and from the internet works upon integrating both into your cloud architecture.

Production-ready outbound connectivity with NAT Gateway and Azure Firewall

For production workloads, Azure recommends separating Azure Firewall and production workloads into a hub and spoke topology. Introducing NAT gateway into this setup is simple and can be done in just a couple short steps. First, deploy Azure Firewall to an Azure Firewall Subnet within the hub virtual network (VNet). Attach NAT gateway to the Azure Firewall Subnet and add up to 16 public IP addresses and you’re done. Once configured, NAT gateway becomes the default route for all outbound traffic from the Azure Firewall Subnet. This means that internet-directed traffic (traffic with the prefix 0.0.0.0/0) routed from the spoke Vnets to the Hub Vnet’s Azure Firewall Subnet will automatically use the NAT gateway to connect outbound. Because NAT gateway is fully managed by Azure, NAT gateway allocates SNAT ports and scales to meet your outbound connectivity needs automatically. No additional configurations are required.

 

Figure: Separate the Azure Firewall from the production workloads in a hub and spoke topology and attach NAT gateway to the Azure Firewall Subnet in the hub virtual network. Once configured, all outbound traffic from your spoke virtual networks is directed through NAT gateway and all return traffic is directed back to the Azure Firewall Public IP to maintain flow symmetry. 

How to set up NAT Gateway with Azure Firewall

To ensure that you have set up your workloads to route to the Azure Firewall Subnet and use NAT gateway for connecting outbound, follow these steps:

Deploy your Firewall to an Azure Firewall Subnet within its own virtual network. This will be the Hub Vnet.
Add NAT gateway to the Azure Firewall Subnet and attach at least one public IP address.
Deploy your workloads to subnets in separate virtual networks. These virtual networks will be the spokes. Create as many spoke Vnets for your workload as needed.
Set up Vnet peering between the hub and spoke Vnets.
Insert a route to the spoke subnets to route 0.0.0.0/0 internet traffic to the Azure Firewall.
Add a network rule to the Firewall policy to allow traffic from the spoke Vnets to the internet.

Refer to this tutorial for step-by step guidance on how to deploy NAT gateway and Azure Firewall in a hub and spoke topology.

Once NAT gateway is deployed to the Azure Firewall Subnet, all outbound traffic is directed through the NAT gateway. Normally, NAT gateway also receives any return traffic. However, in the presence of Azure Firewall, NAT gateway is used for outbound traffic only. All inbound and return traffic is directed through the Azure Firewall in order to ensure traffic flow symmetry.

FAQ

Can NAT gateway be used in a secure hub virtual network architecture with Azure Firewall?

No, NAT gateway is not supported in a secure hub (vWAN) architecture. A hub virtual network architecture as described above must be used instead.

How does NAT gateway work with a zone-redundant Azure Firewall?

NAT gateway is a zonal resource that can provide outbound connectivity from a single zone for a virtual network regardless of whether it used with a zonal or zone-redundant Azure Firewall. To learn more about how to optimize your availability zone deployments with NAT gateway, refer to our last blog.

Benefits of NAT Gateway with Azure Firewall

When it comes to providing outbound connectivity to the internet from cloud architectures using Azure Firewall, look no further than NAT gateway. The benefits of using NAT gateway with Azure Firewall include:

Simple configuration. Attach NAT gateway to the Azure Firewall Subnet in a matter of minutes and start connecting outbound right away. No additional configurations required.
Fully managed by Azure. NAT gateway is fully managed by Azure and automatically scales to meet the demand of your workload.
Requires fewer static public IPs. NAT gateway can be associated with up to 16 static public IP addresses which allows for easy whitelisting at destination endpoints and simpler management of downstream IP filtering rules.
Provides a greater volume of SNAT ports for connecting outbound. NAT gateway can scale to over 1 million SNAT ports when configured to 16 public IP addresses.
Dynamic SNAT port allocation ensures that the full inventory of SNAT ports is available to every virtual machine in your workload. This in turn helps to significantly reduce the risk of SNAT port exhaustion that is common with other SNAT methods.
Secure outbound connectivity. Ensures that no inbound traffic from the internet can reach private resources within your Azure network. All inbound and response traffic is subject to security rules on the Azure Firewall.
Higher data throughput. A standard SKU NAT gateway provides 50 Gbps of data throughput. A standard SKU Azure Firewall provides 30 Gbps of data throughput.

Learn more

For more information on NAT Gateway, Azure Firewall, and how to integrate both into your architectural setup, see:

What is Azure Virtual Network NAT?
Azure Firewall documentation.
Scale SNAT ports with Azure Virtual Network NAT.
Integrate NAT gateway with Azure Firewall in a hub and spoke network.

Quelle: Azure

Microsoft named a Leader in the IDC MarketScape: Worldwide MLOps Platforms 2022 Vendor Assessment

We’re excited to share that Microsoft has been recognized as a Leader in the IDC MarketScape Worldwide Machine Learning Operations (MLOps) Platforms 2022 Vendor Assessment. Microsoft holds a deep understanding of MLOps market trends, strong customer adoption, broad partner ecosystem, and continued product investments in building a differentiated MLOps platform.

The report cited several key strengths including product and business capabilities across the entire machine learning (ML) lifecycle as well as an expansive customer footprint and partner network. 

"Microsoft provides a wide array of enterprise-grade MLOps tools for quality and compliance that can augment any ML development environment on-premises, across clouds, or in a hybrid cloud environment." —IDC MarketScape: Worldwide Machine Learning Operations Platforms 2022 Vendor Assessment

Source: "IDC MarketScape: Worldwide Machine Learning Operations Platforms 2022 Vendor Assessment", By: Kathy Lange, Raghunandhan Kuppuswamy, and David Schubmehl, December 2022, IDC #US48325822.

Do more with less with MLOps on Azure

While machine learning becomes more mainstream across industries, there are many challenges like data governance, security, data compliance, and auditability that companies need to consider when productionizing machine learning models. These organizations look to a unified machine learning platform to manage the entire ML lifecycle to bring models to production faster and at scale, improving operational efficiencies of data science teams with MLOps. At Microsoft, we consider MLOps as a philosophy rather than a product and, as such, our approach incorporates people, processes, and platform to deliver continuous value with fewer resources for machine learning. MLOps bring benefits to organizations by automating repeatable actions, facilitating collaboration across teams, and ensuring full visibility and reproducibility into the end-to-end ML lifecycle.

Azure Machine Learning helps data scientists and ML engineers streamline training, deployment, and management of thousands of models across on-premises, multicloud, and even at the edge using native MLOps capabilities such as model registry, CI/CD pipelines with deep integration with Azure DevOps and GitHub, managed online endpoints for real-time inference, and experiment tracking and lineage with MLflow. PepsiCo has been using MLOps capabilities in Azure Machine Learning to automate model creation and deployment and analyze store customer data to respond to customer demand more efficiently. In fact, they have succeeded in eliminating 4,300 days of manual work a year.

Enhance collaboration across the organization

We believe machine learning is a team sport that requires collaboration across people with a different range of skill sets, such as data scientists, machine learning engineers, and IT admins. One of the key features mentioned in the report is that Azure Machine Learning enables greater collaboration across multiple data science teams through registries. Registries in Azure Machine Learning are organization-wide repositories of machine learning assets: models, environments, and components. Registries make multi-environment MLOps easier by helping teams share and reuse models developed in a different workspace, by a different team without manually copying them over.

Another feature that fosters collaboration between data scientists and business professionals is the Responsible AI dashboard and scorecard, a single interface to help organizations implement Responsible AI in practice effectively and efficiently. The Responsible AI scorecard helps contextualize the model and data health insights with both technical and non-technical audiences, bringing stakeholders along as well as assisting in compliance reviews. Microsoft is committed to accelerating our innovations in Responsible AI.

Microsoft is proud to be recognized as a Leader in the IDC MarketScape Worldwide MLOps Platforms 2022 Vendor Assessment, and we are excited by innovations happening at Microsoft and across the industry that empower data scientists and machine learning engineers to fully operationalize the ML lifecycle. You can read and learn from the report excerpt.

Learn more

Explore other analyst reports for Azure AI.
Read the latest feature announcements from Azure Machine Learning on the Tech Community blog.
Download an excerpt of the IDC Marketscape Worldwide MLOps Platforms 2022 Vendor Assessment to learn why Microsoft is named a Leader.

 

 

About the graphic:

IDC MarketScape vendor analysis model is designed to provide an overview of the competitive fitness of ICT suppliers in a given market. The research methodology utilizes a rigorous scoring methodology based on both qualitative and quantitative criteria that results in a single graphical illustration of each vendor’s position within a given market. The Capabilities score measures vendor product, go-to-market, and business execution in the short term. The Strategy score measures alignment of vendor strategies with customer requirements in a 3-5-year timeframe. Vendor market share is represented by the size of the circles. Vendor year-over-year growth rate relative to the given market is indicated by a plus, neutral, or minus next to the vendor name.
Quelle: Azure

Azure Native NGINXaas makes traffic management secure and simple—now generally available

Continuing Microsoft Azure’s commitment to empower our ISV partners and customers to adopt and modernize their application of choice and run in the cloud, we are excited to announce general availability (GA) of the NGINXaaS offering on Azure.

In facilitating the cloud transformation journey for cloud architects, developers, IT professionals, and business decision makers who are all working towards their digital transformations, we are expanding on our more than a decade of partnership with F5, the company behind NGINX to provide a deeper integration of NGINX into the Azure ecosystem.

NGINX provides load balancing, traffic management, and security tools for users to configure and manage the incredibly complex make-up of the architectures traffic patterns on their cloud and on-premises environments.

“We are excited to expand our Azure ecosystem with the General availability of NGINX for Azure. This strategic partnership with F5 immediately brings together the power of Azure and NGINX’s application delivery expertise to give our developers and customers more native options on Azure.”—Julia Liuson President, Microsoft Developer Division.

Do more with less

Based on inputs from customers in the Open Source world and other users of the NGINX offering, we worked with F5 to simplify the infrastructure management and provide a seamless experience by integrating the deployment, billing, and support of the NGINX solution on the Azure cloud platform, available via the Azure Marketplace.

By taking the management burden away from the user as part of the managed offering, the customer can now focus on the core elements of their business while the custodians of the NGINX and Azure offering bring our strengths to provide a fully managed, secure, and reliable NGINX offering on Azure.

The deep integration into the Azure Control plane also provides another layer of optimization by promoting all the latest relevant features from the Azure Platform to be automatically available to this service.

Deploying and managing load balancer and traffic manager on Azure

The service integrates the NGINX offering into the Azure Control plane. Through this integration, customers can provision a new NGINX service and configure their Azure resources to seamlessly extend workloads to the cloud and deliver secure and high-performance applications using the familiar and trusted load balancing solution. This gives the user consistency in performance and security across their ecosystem via a one-click deployment. In addition, the customers can manage all of the advanced traffic management features they demand, including JSON Web Token authentication and integrated security, to name a few.

Lift and shift from existing deployments

The integrated offering makes it very easy to migrate application delivery from on-premises to Azure cloud. Enterprises and users can now lift and shift apps to Azure cloud seamlessly by bringing their own or existing configurations of NGINX and deploying them from the Azure portal or Marketplace. Users can then configure advanced traffic management and security, leverage state-of-the-art monitoring capabilities, and port custom configurations.

Unified experience

Build end-to-end traffic management solutions with a unified experience. This service gives the user consistency in performance and security across their portfolio of on-premises and Azure cloud apps by using the same load balancing solution and configurations everywhere via the one-click step deployment.

Secure deployments

The ability to control traffic via virtual networks is a critical consideration for our customers. With this integration, users can seamlessly manage configurations between their own virtual network and the NGINX Cloud virtual network via a custom solution leveraging service, Injection. This is further complemented with unified billing for the NGINX service through Azure subscription invoicing.

Getting started with Azure Native NGINX Service:

Discovery and procuring: Azure customers can find the service listed on Azure Marketplace and review the different purchasing plans offered, and purchase it directly with single billing enabled.

Provisioning the NGINX resources: Within several clicks, you can deploy NGINX service in your desired subscription and datacenter regions with your preferred plan.

In Azure Portal experience: Configure the NGINX Networking components:

Configuring logs and metrics: Customers can determine which Azure resource logs and metrics are sent to the NGINX resource.

Learn more

Introducing F5 NGINX for Azure: An Azure Native SaaS Solution for Modern App Delivery.
A Comprehensive Guide to F5 NGINX for Azure: How to get the most out of Azure Native SaaS Solution for Modern App Delivery.
Introducing F5 NGINX for Azure.

Quelle: Azure

Microsoft Azure Load Testing is now generally available

This blog has been coauthored by Ashish Shah, Partner Director of Engineering, Azure Developer Experience.

We are announcing the general availability of Azure Load Testing. Azure Load Testing is a fully managed load-testing service that enables you to generate high-scale load, gain-actionable insights, and ensure the resiliency of your applications and services regardless of where they're hosted. Developers, testers, and engineering teams can use it to optimize application performance, scalability, or capacity.

Get started with Azure Load Testing now, by quickly creating a load test for your web application by using a URL. If you already have load tests leveraging JMeter, you can easily get started by reusing existing Apache JMeter test scripts.

Building resiliency testing into developer workflows

Our goal at Microsoft is to help developers do more with less effort. When performance, scalability, or resiliency issues are identified in production or even close to production they can be extremely difficult and costly to resolve. With Azure Load Testing developers can catch issues closer to code authoring time as part of their developer workflows saving them valuable time and energy.

“As part of our quality shift left initiatives, the Cloud Ecosystem Security teams were able to prevent multiple unique load related bugs from reaching production by gating production builds using Azure Load Testing as part of our CI/CD pipeline. The service teams have also combined the load from Azure Load Testing with fault injection scenarios from Azure Chaos Studio to replicate, root cause and prevent non happy path scenarios that are hard to catch using regular testing frameworks. Along with service resiliency validation, Azure Load Testing has helped uncover the bounds of the distributed system and saved us costs by eliminating unused resources and frameworks.”—Microsoft Cloud Ecosystem Security engineering team

“The Azure Synapse team uses Azure Load Testing to generate different levels of workloads from high concurrency to large input data sequential execution targeting Synapse SQL Serverless endpoints. With the flexibility of JMeter we can start/stop other services within a cluster that can inject different failures, thus truly testing the resiliency of our service.”—Microsoft Azure Synapse engineering team

Pay only for what you need

Optimize your infrastructure while ensuring your application and services are resilient to severe spikes in customer traffic. Leverage Azure Load Testing to optimize your infrastructure before production, planning for the customer traffic you are expecting, paying only for what you need. Then leverage Azure Load Testing to test for unplanned increases in load.

Figure 1: Easily scale load in Azure Load Testing to check the resiliency of your applications and services.

Regression testing

For Azure-based applications, Azure Load Testing collects detailed resource metrics to help you identify performance bottlenecks across your Azure application components. You can automate regression testing by running load tests as part of your continuous integration and continuous deployment (CI/CD) workflow.

 

Figure 2: Build Load Testing into your developer workflow with pass/fail criteria.

Azure-specific insights can help you understand how different load scenarios impact all the parts of your application, and you can compare test results across different load tests to understand behavior changes over time.

Azure Load Testing creates monitoring data using Azure Monitor, including Application insights and Container insights, to capture details from the Azure services. Depending on the type of service, different metrics are available. For example, the number of database reads, the type of HTTP responses, or container resource consumption. Both client-side and server-side metrics are available in the Azure Load Testing dashboard.

Figure 3: Get performance insights across client and Azure service side metrics with Azure Load Testing.

Enable advanced load testing scenarios

For more advanced load testing scenarios, you can create a JMeter-based load test, a popular open-source load and performance tool. For example, your test plan might consist of multiple application requests, or input data and parameters to make the test more dynamic. And if you already have existing JMeter test scripts you can reuse them to create load tests with Azure Load Testing.

Figure 4: Azure Load Testing architecture overview.

What has changed since preview?

Since we debuted Azure Load Testing, we have enabled several new capabilities based on customer feedback.

Quick test creation

Quick test creation of Azure Load Testing with URL. Quick test creation lets you create a load test without a JMeter script, enabling you to set up, run, and test your URL in less than five minutes.

Azure SDK Load Testing Libraries

.NET Azure Load Testing Library
Java Azure Load Testing Library
JavaScript Azure Load Testing Library
Python Azure Load Testing Library

JMeter capabilities

Support for user specified JMeter properties. Support for user-specified JMeter properties, making load tests more configurable.
Splitting input data across multiple test engines. If you're using CSV data in your JMeter script, you can process the input data in parallel across multiple test engines. Azure Load Testing enables you to configure a test to split the data evenly across all engine instances.

Authentication, user-managed identities, and customer-managed keys

Authenticate with Client Certificates. Azure Load Testing now enables you to authenticate application endpoints which require a client certificate.
Test Private Endpoints or applications hosted on-premises. Azure Load Testing enables you to test private application endpoints or applications that you host on-premises.
System assigned and user-assigned managed identities. Azure Load Testing now supports both system-assigned and user-assigned managed identities.
Customer managed keys. Azure Load Testing support for customer-managed keys.

Additional metrics

Additional Client-side metrics for pass/fail criteria. Azure Load Testing enables you to leverage pass/fail criteria metrics including additional client-side metrics of requests per second and latency.
View load engine metrics. Ability to view engine health metrics to understand the performance of the test engine during the run, enabling confidence in the test results and improve test configuration.

Compliance and regional availability

Azure Load Testing is HITRUST certified.
Azure Load Testing Regional availability. Azure Load Testing is now available in 11 regions; Australia East, East Asia, East US, East US2, North Europe, South Central US, Sweden Central, UK South, West Europe, West US2, and West US3.

Get started with Azure Load Testing

You can get started with Azure Load Testing by creating an Azure Load Testing resource in the Azure portal. Check out the Azure Load Testing documentation and create your first load test.

Learn more about pricing details on the Azure Load Testing pricing page.

Watch the new DevOps Lab episode, "What's new in Azure Load Testing?"

Azure Load Testing on DevOps Lab

​

Figure 5: What’s new in Azure Load Testing with April Edwards and Nikita Nallamothu.

Share your feedback

We’d love to hear from you through our feedback forum.
Quelle: Azure

Microsoft Cost Management updates—January 2023

Whether you're a new student, a thriving startup, or the largest enterprise, you have financial constraints and you need to know what you're spending, where it’s being spent, and how to plan for the future. Nobody wants a surprise when it comes to the bill, and this is where Microsoft Cost Management comes in.

We're always looking for ways to learn more about your challenges and how Microsoft Cost Management can help you better understand where you're accruing costs in the cloud, identify and prevent bad spending patterns, and optimize costs to empower you to do more with less. Here are a few of the latest improvements and updates based on your feedback:

Manage your Enterprise Agreement billing account in the Azure portal.
Recent and pinned views in the Cost analysis preview.
Consistent global pricing for the Microsoft Cloud.
Help shape the future of invoice experiences.
Help shape the future of cost management for cloud services.
What's new in Cost Management Labs.
New ways to save money with Microsoft Cloud.
New videos and learning opportunities.
Documentation updates.

Let's dig into the details.

Manage your Enterprise Agreement billing account in the Azure portal

In March, we announced the general availability of the Enterprise Agreement (EA) billing experience in the Azure portal for direct customers working with Microsoft. Now that same experience is generally available for our indirect customers working with partners. All the same EA tools are available from Cost Management and Billing in the Azure portal:

Seamlessly create and manage departments, accounts, and subscriptions.
Manage access to departments, accounts, and subscriptions.
View properties and manage policies, like the ability to view charges and purchase reservations.
View notification contacts for enrollment emails.
View your monthly Azure usage and charges.
Generate and manage API access keys.

Looking beyond account management, you’ll also see new tools to help you monitor and manage costs:

View and download consolidated usage and charges, including options for amortized reservation charges.
Analyze and drill into your costs in the portal or schedule automated exports.
Enable tag inheritance to streamline tag-based cost analysis within your account.
Split shared costs to drive more visibility and accountability throughout the organization with cost allocation.
Configure budgets to get alerted before costs exceed predefined thresholds.

With these updates, EA billing account administrators should start to use the Azure portal for all account management needs. Account management from the EA portal will no longer be available for indirect customers starting on February 20, 2023.

Stay tuned for more updates, including support for indirect partners. To learn more, see EA billing administration on the Azure portal or check out the EA billing administration video series.

Recent and pinned views in the Cost analysis preview

Cost analysis is your tool for interactive analytics and insights. It should be your first stop when you need to explore or get quick answers about your costs. Over the past year, you've seen the addition of new smart views and capabilities, like anomaly detection, that offer more insights and help you understand costs more easily in the Cost analysis preview, but many of you have asked where you should start–in Cost analysis or the Cost analysis preview? Now, you don’t have to choose. The Cost analysis preview lets you decide where to start and remembers which views you use most, helping you jump back in and get the answers you need quicker than ever.

Cost analysis comes with various built-in views that summarize:

Cost of your resources at various levels.
Overarching services spanning all your resources.
Amortized reservation usage.
Cost trends over time.

Cost analysis has two types of views: smart views that offer intelligent insights and more details by default and customizable views you can edit, save, and share to meet your needs. The first time you open the Cost analysis preview, you start with a list of all available cost views.

Smart views open in tabs within the Cost analysis preview, allowing you to switch between views as you investigate issues. To open a second view, select the + to the right of the list of tabs. Customizable views open outside of the tabs in Cost analysis, a customizable view editor.

As you explore the different views, you’ll notice that the Cost analysis preview remembers which views you’ve used in the Recent section. Switch to the All views section to explore all built-in and saved views. If there’s a specific view you’d like quick access to, select Pin to recent from the All views list. You also have quick access rename, subscribe, copy a link to, or delete views from this list.

We encourage you to check out these updates and let us know what you’d like to see next. We’re eager to get your feedback as we continue to evolve the experience for you.

Consistent global pricing for the Microsoft Cloud

Earlier this month, we announced that we are taking several steps to align the pricing of our Microsoft Cloud products (such as, Azure, Microsoft 365) globally, meaning organizations will have consistent pricing reflecting the exchange rate of the local currency to the US dollar (USD). Starting April 1, 2023, pricing for Microsoft Cloud products will be adjusted in the following currencies:

GBP: +9%
DKK, EUR and NOK: +11%
SEK: +15%

In the future, we will assess pricing in local currency as part of a regular twice-a-year cadence, taking into consideration currency fluctuations relative to USD. This will provide increased transparency and predictability globally and move to a pricing model that is most common in our industry.

The Microsoft Cloud continues to be priced competitively, and Microsoft remains deeply committed to the success of its customers and partners. We will continue to invest to enable customers to innovate, consolidate and eliminate operating costs, optimize business performance and efficiency, and provide the foundation for a strong security strategy that customers around the world have come to rely on.

Help shape the future of invoice experiences

Do you view, manage, or pay invoices within the Azure portal or Microsoft 365 admin center? We're exploring new capabilities to improve your invoice experience and would love to get your feedback.

If you are interested in chatting about your experience, please sign up here.

Help shape the future of cost management for cloud services

Are you responsible for purchasing, managing, and optimizing cloud solutions and software for your organization? Does your daily job role involve understanding and monitoring cloud spending, discovering services, acquiring or updating licenses and subscriptions, analyzing resource utilization, or paying invoices?

If so, we’d love to talk to you and learn more about your job role in a 60-minute discussion. Please send an email to CE_UXR@microsoft.com and we will get back to you.

What's new in Cost Management Labs

With Cost Management Labs, you get a sneak peek at what's coming in Microsoft Cost Management and can engage directly with us to share feedback and help us better understand how you use the service, so we can deliver more tuned and optimized experiences. Here are a few features you can see in Cost Management Labs:

New: Remember preview features across sessions.
Select the preview features you're interested in from the Try preview menu and you'll see them enabled by default the next time you visit the portal. No need to enable this option—preview features will be remembered automatically in the preview portal.
New: Customers view for Cloud Solution Provider partners.
View a breakdown of costs by customer and subscription in the Cost analysis preview. Note this view is only available for CSP billing accounts and billing profiles. You can enable this option from the Try preview menu.
New: Total KPI tooltip.
View additional details about what costs are included in the Cost analysis preview. You can enable this option from the Try Preview menu.
Update: Recent and pinned views in the cost analysis preview—Now available in the public portal.
Show all classic and preview views in the cost analysis preview and streamline navigation by prioritizing recently used and pinned views. You can see this in the Cost Management Labs or by opting in using Try Preview.
Recommendations view.
View a summary of cost recommendations that help you optimize your Azure resources in the cost analysis preview. You can opt in using Try Preview.
Forecast in the cost analysis preview.
Show your forecast cost for the period at the top of the cost analysis preview. You can opt in using Try preview.
Group related resources in the cost analysis preview.
Group related resources, like disks under VMs or web apps under App Service plans, by adding a “cm-resource-parent” tag to the child resources with a value of the parent resource ID.
Charts in the cost analysis preview.
View your daily or monthly cost over time in the cost analysis preview. You can opt in using Try Preview.
View cost for your resources.
The cost for your resources is one click away from the resource overview in the preview portal. Just click View cost to quickly jump to the cost of that resource.
Change scope from the menu.
Change scope from the menu for quicker navigation. You can opt-in using Try Preview.

Of course, that's not all. Every change in Microsoft Cost Management is available in Cost Management Labs a week before it's in the full Azure portal or Microsoft 365 admin center. We're eager to hear your thoughts and understand what you'd like to see next. What are you waiting for? Try Cost Management Labs today.

New ways to save money in the Microsoft Cloud

This month I’ll share a few updates spread across the Microsoft Cloud:

General availability: DR secondary free with SQL Server on Azure Virtual Machines.
General availability: Arm-based VMs now available in four additional Azure regions.
Preview: License Geo-redundant Disaster Recovery for SQL Managed Instance for free.
Forrester study finds 228 percent ROI when modernizing applications on Azure PaaS.
Microsoft 365 Basic and more.
Microsoft 365 expands data residency offerings.
Dynamics 365 and Power Platform help you do more with less.

New videos and learning opportunities

Sharing a recent Azure Friday video that does a good job of providing an overview of cost management and optimization for Azure:

Managing, reporting, and reducing your costs in Azure (26 minutes).

Follow the Microsoft Cost Management YouTube channel to stay in the loop with new videos as they’re released and let us know what you'd like to see next.

Want a more guided experience? Start with Control Azure spending and manage bills with Microsoft Cost Management.

Documentation updates

As usual, there were plenty of documentation updates since our last update. Here are a few documents that were updated that you might be interested in:

Group and allocate costs using tag inheritance.
Buy an Azure savings plan.
Microsoft Customer Agreement Azure usage and charges file terms.
Assign roles to Azure Enterprise Agreement service principal names.
Troubleshoot a declined card.
Troubleshoot common Cost Management errors.
30 updates based on your feedback.

Want to keep an eye on all documentation updates? Check out the Cost Management and Billing documentation change history in the azure-docs repository on GitHub. If you see something missing, select Edit at the top of the document and submit a quick pull request. You can also submit a GitHub issue. We welcome and appreciate all contributions!

What's next?

These are just a few of the big updates from the last couple of months. Don't forget to check out the previous Microsoft Cost Management updates. We're always listening and making constant improvements based on your feedback, so please keep the feedback coming.

Follow @MSCostMgmt on Twitter and subscribe to the YouTube channel for updates, tips, and tricks. You can also share ideas and vote up others in the Cost Management feedback forum or join the research panel to participate in a future study and help shape the future of Microsoft Cost Management.
Quelle: Azure

Microsoft named a Leader in the IDC MarketScape: Worldwide General-Purpose Computer Vision AI Software Platform 2022 Vendor Assessment

I am thrilled to announce that Microsoft has been recognized as a Leader in the IDC MarketScape: Worldwide General-purpose Computer Vision AI Software Platform 2022 Vendor Assessment (doc #US49776422, November 2022).

The IDC MarketScape vendor analysis model is designed to provide an overview of the competitive fitness of ICT suppliers in a given market. The research methodology utilizes a rigorous scoring methodology based on both qualitative and quantitative criteria that results in a single graphical illustration of each vendor’s position within a given market. The Capabilities score measures vendor product, go-to-market, and business execution in the short term. The Strategy score measures alignment of vendor strategies with customer requirements in a 3–5-year timeframe. Vendor market share is represented by the size of the icons.

Built on a foundation of deep learning and AI research, Azure Cognitive Services for Vision are designed to be flexible and scalable to meet the needs of a wide range of customers. Our Vision services offer a collection of prebuilt and custom APIs for image and video analysis, text recognition, facial recognition, image captioning, model customization, and more, that developers can easily integrate into their applications. We have been actively contributing to the computer vision research community, further advancing the field and encouraging democratization access to cutting-edge computer vision technology.

Where we excelled

The IDC MarketScape report evaluated Microsoft’s strategies and capabilities and positioned Microsoft in the Leaders Category. I believe this recognition is a testament to Microsoft's commitment to deliver cutting-edge, responsible, and customer-centric AI products to organizations of all sizes and across all verticals.

1) Enabling customers with cutting-edge technology

I believe one of the key strengths that sets us apart is our ability to help customers commercialize cutting-edge technology. For example,

Uber uses vision service’s facial recognition to support security.
H&R Block uses text recognition to transform tax returns.
KPMG uses text, image, and video analysis to help banking customers identify financial risks. 
Territorium to build and deliver a remote proctoring platform and verify student identity.

Recent advancements, such as Dall-E and Florence, have solidified Microsoft’s position as a leader in the industry. Additionally, our OCR and Azure Form Recognizer capabilities have been significantly enhanced, supporting over 300 languages and a broad set of pre-built and custom form models including contracts, invoices, W2s, ID documents, receipts, business cards and upgraded handwriting and customization capabilities. We’ll continue to integrate the latest OpenAI language models with our Vision services to enable deep knowledge queries, further advancing our solutions for customers.

2) Responsible AI

I believe responsible AI is another area where we excel. Microsoft is committed to ethical and transparent AI and has implemented robust processes to ensure our products and services align with industry standards and regulations. Our AI principles, which include fairness, reliability, and safety, guide our development and deployment of AI solutions. We recognize that responsible AI is a journey and are not afraid to make changes based on findings and customer feedback. For example, in June 2022, we announced and implemented the new Responsible AI Standard with additional safeguards and fairness testing and tooling, including a new API that helps developers determine whether an image is suitable for facial recognition, and we shared our plan to phase out emotion detection AI to protect sensitive attributes from being misused—including subjecting people to stereotyping, discrimination, and unfair denial of services.

3) Customer and partner ecosystem

I believe our strong customer and partner ecosystem is also a major differentiator. We have a global network of, partners, service providers, and ISVs that trust and rely on our technology to power their businesses. We work closely with them to understand their unique needs and develop tailored solutions that drive measurable results while supporting their ability to experiment, adopt, and scale Vision service offerings.

As computer vision and AI continue to play an increasingly important role in the digital era, we are well-positioned to help organizations of all sizes and across all verticals leverage the power of computer vision to improve productivity, efficiency, safety, sustainability, and inclusivity. From search engine optimization with image captioning to anomaly detection, facial recognition for identity verification, radiology, or autonomous driving, our platform can support a wide range of use cases and deliver insights that drive business value.

I believe being recognized as a Leader in the IDC MarketScape validates our team's efforts to deliver cutting-edge, responsible, and customer-centric solutions. We will continue to invest in R&D and work closely with our customers and partners to drive the next generation of AI-powered innovation.

Learn more about how Azure AI and Computer Vision lead the way

For more information, please review the IDC MarketScape: General-Purpose Computer Vision AI Software Platforms 2022 Vendor Assessment and discover how our platform can help your organization unleash the power of computer vision. 
Quelle: Azure

Lessons learned optimizing Microsoft’s internal use of Azure

At Microsoft, we learned a lot from moving our internal operations to Microsoft Azure, lessons we use to make our cloud products work better for our customers. As a top user of Azure, we understand our customers’ obstacles and constraints. Just like them, we’re under pressure to "do more with less" in these challenging fiscal times. As such, we're sharing our story of first migrating to Azure, and then optimizing our usage to bring our costs down. Our goal is to help our customers do the same—help them migrate to the cloud, optimize their cloud costs, and strategically invest in projects to boost their growth.

“It’s exciting to talk about our own IT journey, our migration, our optimization, [and] how we’re reinvesting in the different types of solutions that are possible within the cloud, compared to how we did things before." —Pete Apple, Principal TPM Architect at Microsoft

After helping Microsoft migrate to Azure, Apple is now spending his time helping the company get more out of being on the cloud. "We continue to shift [how] we think about the technology, the processes, the ways that we enable our internal customers and our employees to be successful in their day-to-day work and have less friction," Apple says.

Being Microsoft’s customer zero

Many people think we get all our Microsoft technology and solutions for free, but that’s not the case when it comes to Microsoft Azure—we pay full price for it, just like any customer.

“When I talk to customers, they’re always a little bit surprised to hear that we pay the same rates for Azure that they do. We operate with the same licensing, the same pricing, which is really, really important. This informs the role we play as customer zero, when we’re having our own experiences with Azure. We work on improvements with the product group—before our customers can possibly have a bad experience." — Heather Pfluger, General Manager of Infrastructure & Engineering in Microsoft Digital Employee Experience.

Paying full price keeps the pressure on Pfluger’s team to be as efficient as possible with the company’s internal use of Azure. “Like any customer, we must consider competing timelines, optimization goals, and all relevant budgets,” she says. “As customer zero, we not only use Azure cloud services for our business needs—like any other enterprise—but also, we constantly harness our knowledge and insights to envision the right experience in the first place.”

Being Microsoft’s customer zero gives Pfluger and her team the opportunity to give feedback to our product teams, which they use to make improvements that benefit our external customers. We evaluate our solutions to help make sure they meet various global requirements relating to security, privacy, and overall compliance.

Optimizing our cloud costs

Like our customers, we’re feeling pressure to reduce our costs during these challenging economic times. One way we’re doing that is by being efficient with our internal use of Microsoft Azure.

Pfluger was looking for ways to cut costs and says, “The first place we went to was Azure.” The information needed to find ways to trim was easily accessible within her team’s Azure dashboards.

“All of the data was there,” Pfluger says, explaining that there are many optimization levers within Azure that customers can pull. “We saved a tremendous amount of money just by reviewing our Azure spend and by using the tools available to us.”

Our Azure product team is trying to help us, and all our customers get more out of their Azure subscriptions. They’re focused on helping customers understand and forecast their costs, use data to optimize their workloads, use proactive cost-management practices to deploy fixes as quickly as possible, and leverage auditing to identify opportunities to fine-tune their operations even more.

Learn more about optimizing your Azure migration

Read this Azure optimization blog post to get more detail on our internal cloud journey here at Microsoft.
Quelle: Azure