Possibilities of Microsoft Azure
Introduction
Today’s world dives into numbers. Every day we are being overwhelmed by data and at the same time we produce tons of it. Furthermore, relatively recently, we have been able to find out how much knowledge is hidden in this data. Business which is receiving correct information can optimize its portfolio of products or understand the needs of customers. But this knowledge for most people is also really useful as it can improve our own life and ourselves.
The technological process is the main reason of this so-called data fashion. A perfect example of this could be neural networks. The first approach of data analysis using structures similar to the human brain appeared almost 60 years ago as a result of Frank Rosenblatt’s and Charles Wightman’s work. However because of a lack of high computing power the idea has not left the scientific stage. Today the growth of popularity is a result of decrease of infrastructure costs, new computing processors and transferring data into the cloud.
Data science or data analysis is an accelerator of the work of mathematicians, statisticians and IT engineers who, in some cases, bring to life algorithms which have already been known but not used for many years. Thanks to this synergy between different science disciplines, the technological progress can be realized. Data science terminology does not have many publications available on the market so let’s go through this brief to figure out what data science and Microsoft Azure are needed for.
The first question which comes to my mind when I hear about Microsoft Azure is: what exactly is Microsoft Azure? For many, answering this question can be difficult similar to giving the definition of the words ‘smart’ or ‘sustain’. Microsoft Azure is a set of services which can be utilized to assemble solutions. Commonly the Azure environment is a public cloud computing platform that can be used for services such as analytics, virtual computing, data storage, networking and much more. It can also be used to replace local on-premise servers.
It is easy to imagine that all the websites that people come across exist in the invisible virtual world. All the information however is stored in some physical locations processed by computers. Until recently, people have started using remote computing machines for applications and systems. Simply, people have started using electronic devices as portals to this remote information. Microsoft Azure is an open and flexible platform that enables people to build, deploy, analyze and run applications remotely using Microsoft’s data centers. With a rapid expansion of highly advanced electronic devices, a large volume of data has been produced. Azure is a platform which offers outsourcing storage for this data, its processing and managing. In general, data science takes this data for granted as an information source for automatic data processing using platforms like Microsoft Azure. Both the business and scientific worlds have the same approach in this field. Companies and public organizations want to collect all possible data and then, save it, potentially, forever (undefined time). The basic task for data science is to verify different hypotheses based on collected information and then allow people to make practical decisions regarding operations at a company.
There are three main competitors on the market within cloud computing which should be mentioned here. Other companies exist on the market but they are not even able to take over a single piece of the pie from Amazon, Microsoft and Google. The global leader is only one: Amazon is keeping more than 40% of the global cloud sector stepping ahead of Google and Microsoft which each holds around 20% of the market. Broadly, all this would change drastically if a new junior startup invented a new way for connecting data. In fact, the networked world people now inhabit is only possible because of cloud computing. Electronic devices like that smartphone of yours would be just a very expensive thing without the internet. Everything that happens when people use Facebook, Google or social media is accomplished by a combination of servers spread around the world and internet bandwidth. Of course, that doesn’t mean that no computing is being done on a desktop of a computer, but without a connection to the cloud its usefulness would be limited.
Cloud computing was what made services such as Google, the first big web browser, possible. And it was what enabled many companies to move their customers’ services from standalone software to an on-demand service always available and keeping security and privacy as priorities. The truth is, it was Amazon that pushed the technological market into turning cloud computing into a service that anyone can use and rent for a defined time. So data processing and cloud services have gone from being a decentralized business process to one that is centralized and under the control of just a handful of companies.
The Vision of Microsoft Azure
Microsoft Azure is a part of cloud technology platforms. Stanley Morgan estimated that by 2020, Microsoft Azure could be the most popular solution for cloud services beating Amazon that is the current global leader for this kind of service. Because of this, this brief is going to be only about the Azure cloud services.
It’s very handy to have files and personal documents synchronized with online services and be able to review them from any place of the world where the Internet is. However, people can never assume that their data is one hundred percent safe. A good example of this is the second law of computing which states that data doesn’t really exist unless a person has at least two copies of it on different media in different places and deleting one must not affect any of the others. There are solutions and technologies which protect users from similar technical issues but having one more file backup could mean salvation for you.
One of many available services which Microsoft offers to its customers is called. It’s an Internet of Things (IoT) service which allows users to analyze data on devices — at the edge instead of in the cloud. The solution is dedicated to users who want to spend less time transferring data to a cloud and react more quickly to changes in status by moving some parts of the workload to the edge. Azure Edge moves cloud analytics and custom business logic to electronic devices so that business can focus on insights instead of data management. This unreliable benefit gives IT engineers an opportunity to perform some processes such as data cleaning or aggregation locally. IoT Edge modules are components implemented as Docker containers that run your business site at the edge. A user can develop custom modules or packages into solutions that provide insights offline and at the edge.
There is a trend in IoT systems moving towards edge computing but the main reasons why applications perform computing tasks at the edge are privacy, bandwidth, latency and reliability. A great example of this in practice is Google home devices. While the Google Assistant functionality is driven on the cloud, the hot word detection happens locally on the device (saying a keyword e.g. “Hey, Google”). This protects user privacy by avoiding uploads of information until the device knows you are talking to it, but it also eliminates any unneeded costs corresponding with uploading all raw audio to the cloud.
Apart from these terms, reaction time and the ability to operate even when the cloud connection is limited or interrupted are the critical factors especially for industrial automation systems. The advancements in both machine learning and artificial intelligence promise to have big effects on IoT systems in the coming years. The ability to predict correctly customers behaviors and habits will have a huge impact on the market. Based on the electronic devices which collect data, the IoT edge can quickly become a necessary component to the success of IoT as the number of smart devices continues to grow.
But despite the fact that Azure IoT is a brilliant technology, it runs based on the cloud service. But how does it work exactly? Azure cloud service is an example of a platform as a service which means that Microsoft provides a platform allowing to develop, manage and run applications without the complexity and maintaining the infrastructure typically associated with developing and launching an application. A PaaS provider hosts the hardware and software products on its own infrastructure and a user can focus on creating and running code rather than underlying services. A user can install his own software and access it remotely. An Azure Cloud service is typically made available to users via a two-step process. Firstly, a developer uploads an application to the platform storing area. When the system is ready to make an app live, a developer uses the Azure portal to swap staging with production. One of the biggest advantages of this service is an ability to upgrade applications to a new version without disturbing its users. This opportunity is really critical for businesses which have many customers around the world, requiring its clients to upgrade the app. The problem appears when each client has a slightly different system configuration and hardware infrastructure which can generate unneeded technical support tickets and additional work.
Data science and data analytics are fresh topics compared to other technological news. More and more businesses are moving towards services which consist of data collection and data analysis processes. Microsoft isn’t staying behind and offers similar a portfolio of products such as modules of Azure Analysis Services. It provides enterprise grade data models in the cloud offering modelling features to combine data from multiple data sources. As a new user, the Azure portal offers you a set of functions which allows to create a server within minutes and start analyzing data. It is important to ensure that information which is going to be analyzed is understood correctly.
However not many articles are saying something about the dark side of cloud computing which causes carbon emissions. Digital waste has grown exponentially over the last ten years as storing information such as electronic messages (emails), pictures or audio has shifted to the online sphere. An advantage of web services is being able to upload data and leave behind local infrastructure such as discs giving an opportunity to throw all of the information into the digital and virtual global cloud. According to Joseph Romm’s 1999 seminar work, The Internet Economy and Global Warming, indicated that the digital inventories and direct sales using Internet could lead to drastic reductions in energy consumption and greenhouse gas emissions by the year 2010.
However things have gone differently. Every day people produce more and more data which requires a huge amount of server space and energy. As the human digital footprint is growing and there are not so many ideas how to manage, it may be described as, so to speak, digital waste. Today a pro-demand policy where information, software and other products are delivered to an end user via the Internet, it generates an even more scalable number of server farms needed to provide cloud services. A good example of this can be the company MacAfee which reported that the electricity needed to send the trillions of spam emails every year is equal to powering two million households in the US and at the same time generates greenhouse gases that could be produced by at least three million cars.
Google was the first information technology company which noticed this problem and took action to reduce energy consumption at its data centers. The general change of company behaviors, to be more sustainable, gives positive reactions. Using renewable energy sources is recognized as another way to promote efficient computing and operation savings. Over the European continent some server operators — like the German web-space provider Strato — have done some improvements in energy reduction by adopting very high performance energy hardware and software and precise cooling infrastructure systems.
Most technological solutions involve buying more efficient servers and infrastructure but hardware is only part of the existing problem. Another one is software which evolves very quickly. All the information maintained in the cloud in proprietary formats can potentially be inaccessible in the future as the result of data formats which are being changed together with the technology progress.
Summary
There are many different suppliers of cloud computing services but Amazon, Microsoft and Google are the global leaders offering much more benefits for business. Systems built around a cloud-first architecture can already take advantage of services in those companies. To truly scale capabilities in IoT systems, business should push as much machine learning technologies as possible to the edge.
This brief focused on only one company — Microsoft — which offers a product called Microsoft Azure. Azure is a set of cloud services to help businesses and individuals meet their goals within building, managing and deploying applications and data models on a massive, global network. Microsoft offers a platform which consists of different components such as security and management, compute, web and mobile, development services, data analysis, analytics and IoT as well as much more including hybrid operations. Clearly, the speed is a huge advantage of using edge computing technologies.
Cloud computing is a big shift from the traditional way businesses think about IT infrastructure. The main benefits indicated by information technology companies are cost saving, speed, global scale, productivity, high performance and security. Instead of having to invest a lot of money into data centers and servers, companies can save investment capital by using third party resources and only pay for how much the resources are used.
The main disadvantage of cloud computing is the fact that there is a lack of an edge that clearly defines the cloud environment and a personal electronic device. The system can send personal data to the cloud sometimes as a result of unclearly defined rules for users making the entire process seemingly difficult to protect. This is why the European parliament has voted through tougher rules on data protection, aiming at boosting privacy and giving national authorities greater powers to take action against companies that breach these rules. The new data privacy laws comprise the GDPR (General Data Protection Regulation) which manages the use and privacy of EU citizens’ data.
Knowledge is defined as the collection of information about the world and its proper usage. A similar process is being done in services offered by cloud technological companies. They offer portfolios of products which allow users to collect, analyze and manage information that create learning models based on real data. Azure Machine learning service empowers customers to identify end to end machine learning pipelines for problems. However to be able to solve a problem, a deep knowledge about a process should be acquired. Information, in opposite to data, is always standardized with uncertainty as information can be generated in another way which always changes its purpose. The result of this is that although engineers receive access to many different software products, if they do not know core problems and understanding of data, the tools they hold will not help them to solve a problem.
Microsoft Azure is a very useful solution for a company, however the business has to take care of data privacy and delivering engineering knowledge about the processes which are running in the background. One key thing which should be considered by business when it chooses a platform for the cloud computing, is an ability to meet all or almost all business’ requirements and meet all privacy regulations.
—
Written by Lukasz Kudlak
