The Microsoft Azure platform, commonly known as “MS Az”, has been available since 2010 and is one of the fastest growing cloud platforms. Here are a few facts: The market research company Gartner has classified Microsoft with Azure for six years in a row as the world’s leading provider of Infrastructure-as-a-Service (IaaS). Today MS Azure is available in 140 countries – including China. Its infrastructure presence is based on a global network with more than 60 nodes / regions. A number of data centers there ensure stable performance of the platform and its many free services. Four of these nodes are located in Germany, and a total of more than 15 are scattered across Europe.
Azure is GDPR-compliant and certified according to the new international standard ISO / IEC 27701 (Privacy Information Management System PIMS), the requirements of which are much more specific and detailed than those of the GDPR. In addition, Microsoft says that its IoT platform is 100 percent carbon-free. And the software giant employs more than 3,500 experts for research and development in cybersecurity alone, spending one billion US dollars on it.
One of the still major hurdles that companies shy away from using cloud services is their fear of losing control of their data. The Microsoft Cloud counteracts this with a hybrid approach that keeps the platform open on all sides, but still secure: The user can keep all or only certain of his data locally “on premise”. Only processing takes place in the cloud – and that in accordance with the globally recognized ISO / IEC 27701 standard. Microsoft also supports a mix between Azure and other clouds.
To make it as easy as possible for users to manage their multicloud and hybrid cloud activities, Microsoft introduced Azure Arc at the end of 2019. This is a self-service tool that can be used to manage Windows and Linux servers, Kubernetes clusters and Azure data services from any location using a uniform interface. The tool allows the user to access Azure data services and other cloud solutions anywhere.
The local Azure instance is the Azure Stack. With this, services and functions of Azure can be extended to other environments, for example to your own data center or to edge and remote locations. The user can thus create a consistent hybrid cloud environment with identical tools, APIs and processes. There are also various services that function locally as containers or microservices. And when it comes to virtual machines, for example, Azure Backup for Az Hybrid and Az Stack offers additional security.
The so-called cognitive services are among the particularly future-oriented strengths of MS Azure. They make the development of applications in the field of machine learning and artificial intelligence significantly easier. With a simple API call, the developer can embed human skills in his apps: see, hear, speak, search, understand and make decisions.
The Custom Vision tool, for example, supports the visual detection of objects or anomalies. Personalizer can be used to individually place offers for customers and communicate with applications via voice. Such basic functions are already programmed and are available “out of the box”.
Incidentally, Microsoft repeatedly uses its cognitive services for projects that involve the integration of disadvantaged or disabled people. Narrative services, for example, read out texts for blind people, such as a user manual, or an app translates spoken words into the language of the deaf and mute. Another example is another program that performs the exercises of a convalescent who wants to get his musculoskeletal system going again after knee surgery. The app compares his movements with the physiotherapist’s instructions documented in the video and, if necessary, corrects the patient independently by providing feedback on the exercises performed alone.
Microsoft also uses Cognitive Services to track down security problems in and around Azure. According to the provider, the systems detect and stop no less than seven trillion cyberthreats every day.
Data scientists can handle complex machine learning tasks with tools such as Azure Machine Learning Studio, or Az ML Studio for short. Azure provides developers with everything they need – from the calculation machine to the Jupyter notebook. There are also services such as AutoML and MLOps, which make Artificial Intelligence more accessible.
Native cloud applications (NCAs) are becoming increasingly important as they use the possibilities and advantages of the cloud architecture particularly consistently. NCAs consist of many individual microservices that provide the functions and services required for the respective application. Decisive advantage: The microservices can be used extremely flexibly. They are location-independent, require neither specific hardware nor specific operating systems and can be operated on different servers. And they are easy to scale.
MS Azure supports the development and provision of native cloud services in various forms: This can be a container with microservices in an Az Web App, an Az Container Instance or an application that is operated with Az Kubernetes.
Microsoft is also open when promoting NCA frameworks. The company has been providing Dapr (Distributed Application Runtime) as an open source solution since December 2019. The runtime environment supports the developer in creating highly flexible, state-dependent microservice-based applications. Dapr can be used on almost any platform and in any programming language.
Many services that can be used to develop scalable, secure and cost-efficient solutions are also offered by Azure outside of the cloud native approach. And in addition to the services available directly on the platform, the user has access to countless solution offers via the Azure Marketplace. Services such as Az Functions, which quickly provide APIs that react to certain triggers, or the Key Vault to securely store passwords, phrases or certificates.
Software solutions that are implemented in Azure naturally need good storage. In Azure Storage the user finds a very flexible and highly scalable storage. Databases are also well represented, for example with Azure SQL instead of an SQL server on a virtual machine.
For data processing with Az Data Factory, for example, extensive operations can be carried out on data – rule-based, time-controlled or controlled using other triggers. Together with database systems such as Az SQL or NoSQL systems such as Cosmos DB, this results in a coherent range of data processing options.
To bridge the gap back to machine learning and artificial intelligence – there is another useful offer: the Azure Data Lake. It guarantees quick and easy availability of data from a wide variety of sources.
Azure is now the IoT platform used in many projects. For example, when networking wind turbines with the cloud platform. The data collected can be used, among other things, to make predictions about the behavior of the system transmission. And the information a
vailable via the cloud about the transmission behavior in real operation enables predictive intervention. The data make it possible, for example, to provide machine spare parts in advance and to plan the replacement before damage has occurred. In this way, the service life of the gearboxes can be extended and optimal control of wind farms can be guaranteed.
The entrance and exit gate for IoT communication is the Azure IoT Hub. There is hardly a scenario that would not be feasible in this way – right up to the edge, i.e. the “non-connected” machine that can be connected and managed via Az IoT Edge, for example. Whether with a hybrid approach, with container or Azure native – the Azure IoT Services offer highly available and strongly secured services that can be easily scaled.
The data flow, which is important for IoT digitization projects, goes for Azure via the Azure Service Bus or the Az Event Hub. In order to reliably process the events of an IoT device, Azure also offers the Az Event Grid. This service package is suitable for carrying out IoT digitization projects.
If you are preparing an IoT project, Azure DevOps is an agile software planning and development tool that makes projects effective and transparent for the company and development teams.
Another selection of development tools for the implementation of digitization projects is the Azure Marketplace. Here the software developer can find certified solution modules or complete solutions from other providers with which he can start his project.
Incidentally, you can easily start “out of the box” with Azure IoT Central and implement your solution idea on a small scale. Here the developer will find: rules, roles and rights concept, user interface adaptation, AI and data analytics through to alarms. With little effort, it is possible to find out how good the chances are that the project will probably be successful.
Azure provides suitable tools for message or data communication within an IoT application with the Az Event Grid, Az Service Bus and Event Hubs.
Hardly any other platform is as well documented as Azure. It is also helpful that a very active and open community has developed around Azure that is happy to share its knowledge. The libraries related to Azure also provide support. Programmers can find helpful instructions here. Two examples: Phyton and .NET. The development language plays almost no role here. Support for .NET is still the best, but Java is very close to this level of support, and other languages such as Python are also very well supported.
With the Azure price calculator, the developer can calculate the services that he intends to provide for the project before starting his project. In addition, the platform offers ways of scaling services in such a way that they generate a minimum of costs but a maximum of benefits. If, for example, a virtual machine is no longer required within a defined period of time (“idled”), it is shut down. The VM only incurs storage costs – nothing more.
Whether a company, or its developers, opts for MS Azure or Amazon Web Services (AWS) as an alternative for their cloud applications depends heavily on the business model and the type of application. It is not surprising that AWS is considered a good platform for highly scalable web applications. After all, the Amazon business model lives from services that are used by hundreds of millions of users worldwide.
On the other hand, Azure is very strong due to its wide range and the distinct possibility of implementing hybrid or cloud-native projects. In addition, there is the openness towards Open Source and Linux, the high innovation potential of the services and Microsoft’s ongoing investment in security.
As far as costs are concerned, a comparison is out of the question, if only because countless parameters play a role. In any case, it is helpful for beginners to first use a test version and estimate the expected costs with the Azure price calculator or the AWS Pricing Calculator before starting an operational project. (bw)