How PaaS are making app development faster and more affordable? With PaaS ("platform as a service") cloud services, even companies without deep IT expertise can quickly build their own business applications. We analyze how these services work and what they mean for business
PaaS is the most complex and interesting segment of cloud services. It occupies a special position between IaaS ("infrastructure as a service") and SaaS ("software as a service"). Explaining what PaaS is and how such services work is a non-trivial task in itself.
The principles of IaaS and SaaS are generally simple and straightforward. IaaS are processors, RAM, data storage, only not physical, but virtual, created in the cloud provider's data centre.
With SaaS it's even simpler. For example, all web applications – mail, messengers, video/audio players, navigators – are SaaS. The application runs on a remote server, and the result is displayed on the user's computer, smartphone, or tablet.
PaaS services work somewhat differently and are more comprehensive. Under the PaaS model, the provider provides access to a cloud environment in which to create, test, scale and update their own applications. The PaaS user gets all the tools needed for development – the operating system, middleware, databases, and more – out-of-the-box.
Prior to PaaS, IT teams had to select, purchase, configure, integrate, and maintain an entire suite of products on their own. It was time-consuming, complex and expensive work. Only when they were done could developers start building applications themselves.
Now, almost everything they need can be found on a single platform. The experts only have to write the code and test the new application.
One of the world's leading providers of cloud-based SaaS solutions, SaleForce, describes its Force.com platform as Development as a Service, emphasising its usefulness for development. There is logic in this: when choosing between the basic options of cloud services, PaaS is really best suited for developing one's own software products.
In the case of IaaS, the user only gets access to the infrastructure – storage, networks, server and other computing power. It requires a lot of preparatory work before you can start building a product. SaaS is an off-the-shelf application that may be suitable for a particular business problem, but… it may not.
If a company were to cook a new meal rather than develop an application, it would get a well-equipped kitchen with a full set of utensils, products and spices under the PaaS model. Under IaaS, he would get a room with a refrigerator, a cooker and an oven. And in SaaS, he would get a cooked meal.
Thus, PaaS is a service for those who have decided to "cook" themselves, but do not want to delve too deeply into the technical details and spend time, labour and financial resources on creating a development environment.
At the end of last year, Gartner analysts counted more than 360 operators in the global PaaS market offering over 550 different cloud platforms. Experts believe that by 2022, this segment will grow by more than a third compared to last year's figures and will reach $34 billion.
PaaS services today are provided by both major cloud market players – Amazon Web Services and Microsoft Azure – and national cloud service providers.
In addition to basic cloud services, there are services such as Serverless Computing (Functiongraph), Message Broker (Distributed Message Service), Application Orchestration Service, Microservice Support (Service Stage) and many other PaaS tools. This makes it possible to build products of any level of complexity on the basis of the platform – from a simple backup system to "smart" e-commerce and solutions using artificial intelligence. It is possible, for example, to run a "smart" and reliable online shop that will withstand any Black Friday surge and provide the retailer with full business intelligence. And it is possible to use cloud PaaS services to work with big data, as Mediascope has done.
PaaS are most actively used by companies with a high level of IT expertise, as well as by businesses operating in highly competitive markets – from retail to the HR industry and other service companies. According to SberCloud experts, one of the latest trends in the use of cloud solutions has been the increased use of services for automation of deployment, scaling and management of applications based on containerization technology and microservice architecture.
What the complete software environment that PaaS users receive consists of:
These include, for example, a source code editor, debuggers and compilers. The specific set of tools depends on the platform.
The end user does not interact with this software, but this software is necessary for developing new applications. Recently, this area has increasingly been identified as a separate segment, Middleware as a Service (MWaaS). The service usually includes an application server and integrated security features.
This service provides the user with access to databases of any type. The provider provides database administration and maintenance, offloading the workload from the company's IT staff. This service is one of the most popular on the market. It is also sometimes singled out as a separate segment of platform services – Data Base as a Service (DBaaS).
Because a PaaS provider takes care of most of the tasks and provides out-of-the-box tools, using such services dramatically simplifies the process and speeds up development. On a cloud platform, the IT team can almost simultaneously test different configurations to verify compatibility and performance.
As a result, time-to-market is dramatically reduced, enabling the company to bring new services and products to market as quickly as possible. This is important not only for the IT industry, but practically for any business.
Another important benefit is cost optimisation. For one thing, companies do not need to invest in their own infrastructure. Platform users receive the service for a relatively small subscription fee, and many providers offer a pay-as-you-go model. As a result, cost reductions, particularly in the start-up phase, can be as high as 90%.
On the other hand, an off-the-shelf software environment and automation of development processes reduces the number of man-hours required to create products. As a rule, PaaS saves up to 50-70% of IT team time.
But this cost reduction doesn't limit what users can do. On the contrary, with out-of-the-box tools in hand, companies can focus on the product and devote more time to its development. And, just as importantly, it doesn't have to be done at the head office. Cloud platforms allow you to work remotely (which is especially important right now), involve distributed teams and gather developers from all over the world.
As new business needs arise, cloud platforms are being filled with new options and features. "The trends we are seeing in PaaS both reflect and drive the ongoing transformation of cloud computing and digital business," says Efim Natis, research vice president at Gartner.
This means that the toolset available on platforms will expand to include the most in-demand and emerging technologies – big data, neural networks, artificial intelligence and machine learning. Companies will be able to use them without having to dive into the intricacies of programming and administration.
In public clouds, thousands of different companies use the infrastructure and the provider takes care of the management. This approach is easier and cheaper for the customer and is increasingly being used for application development and testing.
In addition, major cloud service providers are now offering customers powerful cloud computing clusters for artificial intelligence. They are needed to quickly process large and complex data sets, quickly train AI models and run various machine learning and deep learning solutions based on them. Thus, at the end of 2019, SberCloud launched Russia's most powerful supercomputer, Kristofari, and the AI Cloud platform. Two important services – Model Training and Model Inference – are implemented on the platform. The first enables training of machine learning and deep learning models on Kristofari. The second allows for the deployment and use of artificial intelligence models.
In essence, they combine IaaS (supercomputer capacity) and PaaS (familiar to data scientists development environments and tools). And now any company has access to supercomputing and AI-based development that used to be available only to large corporations.