Contents
The development of the market and technology sets tasks for companies that are difficult and expensive to perform on a traditional IT infrastructure. Understanding what new solutions and tools are needed today
About the expert: Alexander Tugov, Selectel Services Development Director.
Challenge: build an ecosystem
Large companies – primarily in IT and finance – are actively engaged in the creation of ecosystems. For many players, this is a natural result of growth and development.
The business starts with one successful service, then launches new products and services. When a company has five or six successful services, there is a need for integration. Then there are ecosystems that can generate additional value for the user – for example, more favorable conditions, ease of registration and identification. In particular, Apple, which created one of the first and most powerful b2c ecosystems, recently released the Apple One subscription package. Using it, you can connect to any user services from any personal devices.
The ecosystem approach gives companies the opportunity to sell more services to the same user. Moreover, ecosystems allow you to move from individual services to solving complex user scenarios. But such an approach requires appropriate technical resources and components.
What infrastructure is needed
Providers’ business models are also adapting to the new demand and moving towards ecosystems. We at Selectel started with classic data center services and dedicated servers, then cloud storage, virtual private cloud and other cloud services appeared. At the first stages, we did not connect them in any way, so as not to slow down the development and their launch on the market. But when the products matured to a mature state, the company began to focus on an integrated approach.
Now Selectel does not just offer dozens of different services on the showcase, but selects and integrates them to solve the client’s problem. And if it needs a dedicated server, a piece of the cloud and computing power, the client receives them in the form of a ready-made set of tools.
For example, based on a cloud storage or platform, you can combine different systems for managing roles and access rights. Suppose a client has several lists of users in storage for different services, and you need to differentiate rights between these groups. Previously, it was required to first manage one part of the users, then another. Moreover, this had to be done in different parts of the control panel or even different applications. But the integration of systems that we are currently carrying out will allow us to manage all user lists and access rights from one point.
Challenge: combine technologies and levels of protection
Our client, who has a website for selecting apartments for purchase and rent, was planning to launch a new service for selecting a mortgage. But unlike the main site, it provides for working with personal data: scans of passports and other important documents are loaded into the system. This means that the resources for the new service must have an increased level of protection and be certified in accordance with the law.
Many companies face similar situations. Especially now, when products and services are transferred online, and in order to provide them, personal data of the client is needed.
What infrastructure is needed
A mobile app or user portal can rely on multiple systems and solutions at the same time. For example, a hybrid of private and public clouds or a combination of on-premises solutions with cloud infrastructure.
Thus, the mentioned client company placed its mortgage service in the secure perimeter of the Selectel private cloud. We connected these resources with the old information system on the company’s dedicated servers.
In fact, the site has two IT systems under the hood, built on completely different technologies and united by the provider. But users do not notice the transition when using both services in a single application or site.
Challenge: Accelerate product development
When launching new products and projects, companies strive to immediately make them flexible and scalable. But most importantly, they are interested in the launch process being quick and as economical as possible.
But if you create products and solutions completely from scratch, you will need time and an expensive development team with deep expertise in IT. Therefore, tools are required that reduce time and cost.
What infrastructure is needed
Platform services and elements of the development environment allow you to speed up and simplify the launch process. For example, Kubernetes technology that helps automate the deployment and management of containers. In fact, these are ready-made software units into which the application is “packed” along with the code, launch environment, libraries and settings. Today it is one of the most popular requests among our clients.
In addition, databases as a service (DBaaS) are in high demand. These are familiar databases, but completely managed by the provider. He is responsible for scaling, fault tolerance and backups. The client only needs to connect and work with data
Such tools greatly reduce development time and, as a result, time to market. And the company’s IT staff can focus on the product rather than building and maintaining the software environment.
At the same time, these services are integrated with all other services. For example, you can use dedicated servers for running applications, and move databases to DBaaS. Or leave legacy applications in the cloud based on Vmware, and deploy new ones in Kubernetes in the cloud platform.
Challenge: implement machine learning
Now business uses machine learning and artificial intelligence technologies mainly for voice assistants and chatbots. They facilitate communication with users, but hardly bring tangible benefits to the traditional business.
However, companies are gradually starting to use machine learning and neural networks to solve specific business problems — cost optimization, redistribution of financial flows, debugging logistics.
Most likely, such enterprise solutions will evolve in the same way as user programs. For example, in the latest version of Photoshop, machine learning technologies allow you to instantly change the color of the sky, lighting, palette, and even shadows in a photo. Previously, this was done by experienced editors and illustrators in a few hours. Now an ordinary user can do the same, and in just a couple of clicks.
Corporate business applications with neural networks will also simplify processes and free professionals from routine tasks. Moreover, companies will be able to collect and extract more data about the work of enterprises and departments. And if earlier this data did not make sense, then machine learning will help to analyze it and find points for optimization.
What infrastructure is needed
The use of neural networks and machine learning places increased demands on computing power. The more complex algorithms are used, the more computing resources are needed. The training of neural networks takes a lot of resources.
A standard server, which considers the conditional “arithmetic mean” for several indicators, will cost about 200 thousand rubles. If you put several GPU accelerators in it, which allow you to quickly solve machine learning problems, then the cost will already vary from half a million to millions of rubles. Such expenses can seriously affect the company’s budget.
But learning tasks are periodic. If a company needs to train a neural network within one to two weeks, using the cloud for these purposes will cost 10-20 thousand rubles.
With this approach, it is not necessary to build up your own IT competencies and develop an IT department that will only consume company resources. Instead, you can focus on the tasks and areas that bring money here and now.
What does all this mean for the market?
The considered examples confirm that most of the tasks facing businesses today require an integrated approach to the IT infrastructure. That is why we at Selectel have chosen the path towards ecosystems and shifted the focus from individual products to solving any customer problems.
Until this transition is completed. But the first projects show that companies need complex solutions. Today, dozens and even hundreds of different products are presented on the platforms and showcases of providers. Clients do not always have the desire and opportunity to study their subtleties and look for the most suitable options. The task of the provider is to take this headache upon themselves and assemble the ideal constructor for the needs of a particular business.
Both parties are interested in such an approach. It allows client companies to quickly receive the necessary services and not waste resources on the selection of solutions. And it is more efficient for providers to sell their products. Moreover, both those and others can be sure that all the chosen solutions fit well with each other and provide synergy, and not constant failures.
Subscribe also to the Trends Telegram channel and stay up to date with current trends and forecasts about the future of technology, economics, education and innovation.