MWS Cloud

Main article:

• What is IaaS
• Virtualization. Classification and applications
• Virtual Desktop Infrastructure (VDI)

The MWS cloud is a fault-tolerant IT infrastructure based on software and hardware from global manufacturers and located in MTS partner and own data centers that meet the TIER III reliability level according to the Uptime Institute TIA-942 standard and methodology.

2025

Capacity expansion of virtual infrastructure with GPU by 1.5 times

MWS Cloud in the first half of 2025 received a 1.5-fold expansion of virtual infrastructure capacity with GPU. The virtual infrastructure was expanded in two Moscow data centers - in Avantage and GreenBushDC, as well as in the data center in St. Petersburg. GPU power is available in the Pay as you go format.

The virtual infrastructure with GPU allows you to train and infer any ML, LLM and CV models. On top of it, you can deploy an ML platform or a platform for the MWS Cloud interface. And payment takes place according to the Pay as you go model - only for the resources that are consumed.

The capacity is for developers, data Scientists, and companies that:

• They train complex ML and LLM models (including generative neural networks), launch the inference of highly loaded commercial products, develop computer vision (CV) for recognition and analysis systems, and use resource-intensive computing for rendering, simulations and working with big data.

We see a steady increase in demand for virtual infrastructure with GPUs. Since the beginning of the year, its consumption in the MWS Cloud has increased almost 1.7 times. In this regard, we are striving to proportionately expand our cloud capacity, - said Aleksei Kuznetsov, director of new products at MWS Cloud.

Inference Valve launch

MWS Cloud has launched a platform for the introduction and operation of artificial intelligence models - Inference Valve. MTS Web Services, MWS (MTS Web Services, MVS) announced this on August 12, 2025. It allows you to produce any trained ML models, large language models (LLM) and computer vision models (CV). Using the platform, they can be deployed on the infrastructure, connected to the company's IT systems through standard APIs, scaled, and updated and monitored. The platform allows you to significantly reduce the time for introducing models into business processes and reduce the operating load on teams by up to 70% by automating the routine - from packaging the model to its maintenance. After the launch of the MWS Cloud platform, it entered the platform market for inference AI models with a volume of more than 15 billion rubles.

With the platform, CV models can automatically analyze video by finding specific objects, scenes, and actions, or classify medical images. ML and LLM models will be able to generate and structure texts, analyze data, create a description of it, predict marketing and sales indicators, and answer frequently asked questions in ^[1]. You can also deploy voice models on the platform for speech synthesis and analysis, as well as introducing them into call centers.

Companies can deploy both their own trained AI models and use ready-to-use open source models. In addition, Inference Valve integrates with the ML platform and continuous development (CI/CD) tools, which allows you to build a seamless process from AI training to its implementation in business processes and continuous update.

The platform is available in a private cloud on the MWS Cloud infrastructure, on-prem on customer servers, as well as as part of hardware and software complexes (PAC) in a closed loop, including modes with limited access to external networks.

After starting the cluster, the customer's specialists load the model artifacts (for example, ONNX, TorchScript) into the platform, after which it automatically forms the service container and publishes the endpoint. The platform supports simultaneous operation with several models at once with allocation of quotas for computing resources, version control, routing traffic between versions and scaling under load on both GPU and CPU.

Inference Valve also provides latency and bandwidth metrics, availability monitoring, alerts, and dashboards; quality telemetry is available, including data and model drift tracking, target metrics control, and degradation notifications. Integration with observability systems (Prometheus/Grafana) and query logging makes it easier to audit and analyze incidents.

{{quote 'Without the active introduction of artificial intelligence, it is almost impossible to count on market leadership. In MWS Cloud alone, the demand for AI infrastructure doubled over the year. At the same time, it is not enough to simply train the AI model. It still needs to be integrated into the business processes of the company, to establish its uninterrupted operation, to monitor the success of its work. Inference Valve allows you to do this out of the box with a few simple commands. With its help, LLM and CV models integrate tens of times faster with the company's IT systems, the operating load on ML teams during the operation of models is reduced by 70%, the degree of CI/CD automation increases by more than a third, and GPU costs are reduced by more than 15%, - said MWS Cloud CEO, Executive Director of MTS Web Services Igor Zarubinsky. }}

1C hosting service

MWS Cloud has launched hosting 1C an in-cloud application service. As part of the service, the company will deploy applications for customers, 1C provide users with access to it and take over its configuration and support. With the release of this service, MWS Cloud will be able to provide the entire list of services related to 1C. This was MWS announced on August 7, 2026.

MWS Cloud will deploy 1C in a virtual infrastructure with 3.25 GHz processors and a large amount of RAM. As part of the service, the company will also take over the configuration, support and monitoring of the virtual infrastructure, the 1C platform and the DBMS, and, if necessary, the 1C client applications themselves. MWS Cloud specialists can also migrate existing 1C services to the cloud.

The service is deployed in its own data centers MWS Cloud of the Tier III fault tolerance level, in a virtual infrastructure certified according to the requirements of the 152-FZ for working with personal data.

The Cloud 1C service has entered the line of 1C services provided by MWS Cloud. In addition to it, the company can provide separate licenses: they can be either bought or rented. In addition, MWS Cloud can provide "Secure 1C Hosting," which is additionally protected by 7 information security factors. Among them: antiDDoS, WAF, SOC and backups.

The 1C Cloud is an easy and affordable way to get started with 1C in a virtual infrastructure. Companies will no longer have to wait for the delivery of equipment or configure and support applications themselves. Applications are deployed in minutes, and migrating to the cloud as part of a workday. All this happens seamlessly for business. In fact, companies receive a set of 1C services and the cloud simply at the price of virtual infrastructure, "said Mikhail Tutaev, MWS Cloud Product Director.

Capacity increase by 3.3 times

MWS Cloud, included in, MTS Web Services increased virtual the infrastructure capacity by a third in two: Moscow data centers"" and Avantage. GreenBushDC The company also launched a second availability zone in, Novosibirsk making the total capacity of the MWS Cloud in Siberia 3.3 times higher. MWS announced this on July 30, 2025.

Data centers have deployed servers with high-frequency 3.25 GHz processors and a large amount of RAM, which are best suited for hosting databases and SAP and 1C solutions. On top of the virtual infrastructure, it will also be possible to deploy any other cloud services of the company, such as services for working with MWS Data data, business applications, including Russian software or services for developers.

The new capacity will allow companies from the European part of Russia and Siberia to increase the consumption of cloud capacity without investing in their own infrastructure and without delays in the supply and installation of equipment.

Cloud power is available through the MWS Cloud Platform management console. You can access them by charging Pay-as-you-go, which allows you to pay only for the resources that are used.

A new availability zone in Novosibirsk was launched in the first modular data center of MTS beyond the Urals. The data center is certified according to TIER III requirements, which indicates a high level of infrastructure reliability. The capacity of the data center is 250 server racks. Due to the fact that there are two cloud accessibility zones in Novosibirsk, MWS Cloud has the opportunity to make fault-tolerant solutions in the very center of the country. When, for example, an IT system or database is deployed in one data center, and their backups or disaster recovery services in the cloud are deployed in another.

"At the beginning of the year, the consumption of our cloud services increased by more than a third. We strive to increase the capacity of our infrastructure in proportion to the growth of the market so that our customers can always get unlimited opportunities to scale their resources. At the same time, we are launching new availability zones not only in Moscow and St. Petersburg, but also in more remote regions so that our customers can transfer their data to the infrastructure as close as possible to their offices and production facilities, "said the director of the grocery stream "Computing, Storage, Networks and Media" MWS Cloud Pavel Bragin.

Launch of the virtual infrastructure segment in the modular data center of MTS Fedorovsky

MTS Web Services (MWS) on January 20, 2025 announced the launch of the virtual infrastructure segment in its own modular data center "Fedorovsky" in the Leningrad Region. This cluster is designed to accommodate IT systems of companies that require a large amount of computing power. Also, all key MWS cloud services will be provided on the basis of the cluster. Total cloud capacity in the region grew by more than a third.

The cluster has advanced computing characteristics. It will be possible to deploy any high-load business systems: from enterprise management systems to sites with millions of visitors daily. Based on the computing infrastructure, all key MWS cloud services, such as backup, object storage, virtual desktops, or office software, can be deployed in the future. The entire infrastructure will be deployed in the Fedorovsky modular data center, corresponding to the Tier-III fault tolerance level.

The cloud allows companies to save time placing orders for equipment, its logistics and installation, which means speeding up the launch of new projects. The virtual infrastructure rises in minutes, the business has the opportunity to increase or decrease its volume depending on the need and more flexibly manage its spending in general. This means that if an enterprise uses less computing power than it calculated, then its servers are not idle, and if more, there is no need to wait for new supplies. This leads to an increase in demand for virtual infrastructure from Russian companies. In 2024, the domestic cloud market grew by 32%, according to MWS research. To meet growing customer demand, we regularly expand our virtual infrastructure. The cluster in Fedorovsky has already become the 15th zone of availability of MWS cloud services, - said Mikhail Tutaev, Product Director of MTS Web Services.

Launch a service for the safe operation of AI with personal data

MTS Web Services (MWS) on January 14, 2025 announced the launch of a virtual infrastructure segment with a GPU certified for compliance with 152-FZ requirements for the first level of security. The service will allow companies to safely work in a cloud with artificial intelligence, the training of which requires personal data (PD). It will include all the tools necessary for AI training - from a virtual infrastructure prepared for working with personal data to special software for interacting with ML models.

Any personal data can be transferred to the virtual infrastructure segment with the GPU: from contact information to biometric data and special sets of PDs, such as nationality, philosophical beliefs or medical data. They can be used to train artificial intelligence.

ML models will be able to help doctors diagnose and conduct research based on patient data, companies will be able to use AI to process information about buyers and launch chatbots, which, among other things, will be able to collect PD.

To train artificial intelligence on personal data in the cloud, companies will be able to both use the MWS ML platform already ready for operation, and deploy their own AI training software, as well as their own ML models, in a virtual infrastructure.

Any companies working with personal data will be able to use the service. However, it will be most in demand in the field, health care, and retail. e-commerce financial industries

The importance of artificial intelligence is growing every year. In the past year alone, sales of capacity for his training have nearly quadrupled. The number of personal data collected by companies is also increasing. Analyzing them with AI can help streamline enterprise business processes and significantly increase profits. At the same time, in order to work with PD, it is necessary to comply with the requirements for their safe processing. This service will allow businesses not to engage in data protection on their own by transferring these functions to the cloud. Within one window, enterprises will receive all the tools necessary to work with AI: from computational power protected for working with PD to specialized software for training ML models, - said Aleksei Kuznetsov, director of new cloud products at MTS Web Services.

2024

Start Protected 1C Hosting

MTS Web Services (MWS) On December 5, 2024, it announced that a cloud comprehensive protected service had become available in. 1Chosting Customers will now be able to place the 1C configuration on the high-performance servers with the maximum. data protection

The secure 1C hosting service will complement the range of services for ensuring the continuity of 1C systems provided to customers from the cloud. Depending on the needs of the business, a choice of 1C support functions is available - the sale and lease of licenses, as well as support by certified specialists. To ensure maximum security and performance, remote access to the information base is carried out through a web/thin client, as well as over an encrypted communication channel.

The security features built into the service ensure security, data prevent confidential information leaks , and eliminate compliance issues. 7 tools are available to users from information security trusted vendors: isolated resource loop, (multivariate authentication MFA), protection of web applications from hacking (WAF), access to the service through secure communication channels (GOST), VPN antivirus system backup and round-the-clock monitoring, as well as incident response by a information security team of specialized specialists.

Continuous business processes and uninterrupted employee access to 1C applications is ensured by a fault-tolerant infrastructure located in reliable, data centers Level Tier III-compliant and fully recoverable backup systems.

Placing 1C in a reliable cloud space allows you to optimize the cost of maintaining and protecting IT equipment, and by integrating with information security services, significantly increase the level of data security and organize secure remote access for employees to all the information they need. The flexible scalability of the solution allows you to quickly increase capacity if necessary. The service is compatible with all operating systems, so you can work with it from any device, - said Mikhail Tutaev, Product Director of MTS Web Services.

The possibility of private VDI installations using Russian virtualization technologies and OS

MTS Web Services (MWS) on November 20, 2024 announced that the private installation service of Virtual Desktops (VDI) using Russian virtualization technologies and operating systems is now available in the cloud.

The private installation of Virtual Desktops allows you to create your own secure VDI infrastructure with unlimited scalability based on dedicated servers. Installation on servers is carried out in accordance with individually selected equipment configurations for the specific needs of the customer. The personal VDI cluster can be administered both independently and by the provider.

If necessary, a private VDI installation integrates quickly and cost-effectively with public cloud resources, which is relevant, for example, to support peak loads, organize backups, or deploy test environments.

The dedicated infrastructure has a completely closed circuit and is located in reliable data centers on the territory of the Russian Federation that comply with the Tier III Certification of Operational Sustainability standard, certified according to information security requirements for state information systems and in accordance with the 152-FZ on personal data.

to storage Access control allows you to ensure the highest requirements for and processing information with the ability to restrict or completely prohibit access to provider administrators. Additionally, on the basis of the dedicated equipment, you can create a site for backup and recovery data , which guarantees the disaster tolerance of the infrastructure.

In the last year, we have seen an increase in the interest of Russian companies in Russian VDI technologies. The creation of private VDI installations using Russian technologies in the MWS infrastructure will allow large customers to create an independent, secure and stable virtual workplace infrastructure.As part of the service, MWS experts select the necessary equipment and system software in accordance with customer requirements, and complete configuration of virtual desktop management systems and further support, - said Sergey Romanenko, Head of VDI MTS Web Services.

Launch of BI platforms

MTS Web Services (MWS) on October 3, 2024 announced the launch of cloud systems for analyzing and visualizing data of any complexity (BI platforms), which collect, store and visually present data to identify trends and important business metrics.

BI platforms will complement the MWS cloud ecosystem, which already allows you to securely store data of all sizes and distribute content to millions of users, as well as train ML models, manage their lifecycle and conduct joint experiments.

Three business intelligence systems will be available in the MWS cloud: Glarus BI - flexible platform for companies with unique business processes and special individual needs, Modus BI - a solution for medium and small businesses that need a simple and intuitive tool with extensive integration capabilities, as well as a platform for the Enterprise segment - Luxms BI, which provides powerful analysis and visualization tools to help you monitor, analyze performance indicators and prepare interactive reports to address specific business challenges. As of October 2024, Glarus BI is already available to users for work, the rest will appear in the cloud before the end of 2024.

The platforms meet all the requirements of the import substitution policy, are included in the register of domestic software, are compatible with Russian OS, have documentation and technical support in Russian.

BI solutions will help companies analyze current business indicators and manage efficiency, identifying barriers and growth points in real time, to quickly adapt the development strategy to the needs of the business. All systems use a modern technological stack, and their implementation in the cloud guarantees companies rapid implementation, high speed of operation, flexibility and scalability, - said Mikhail Tutaev, Product Director of MTS Web Services.

Launch a dedicated cloud segment on domestic software

MTS Web Services (MWS), part of the MTS group, on September 9, 2024 announced the launch of a dedicated cloud segment on Russian software and equipment. Learn more here.

Launch two virtual infrastructure clusters with graphics accelerators

MTS Web Services (MWS), part of the group, MTS on August 28, 2024, it announced the launch of two infrastructure clusters virtual with graphics accelerators (GPUs) DPCs at Fedorovsky St. Petersburg GreenBushDC in and. These To Moscow segments will speed up and simplify implementation artificial intelligence the Russian by companies. After the launch of these segments, MWS GPU power increased by 40%.

Graphics accelerators are needed to work with artificial intelligence models. The launch of segments will allow more enterprises to simplify and speed up training and performance of their own AI- and ML-models by placing them cloud in MWS. The infrastructure is suitable for working with any AI models in everyone. And industries economies you can access it from anywhere in the country.

Also, by the end of 2024, GPU-infrastructure will be deployed: a platform for the development, training and deployment of ML-models, ready-made AI models for different industries and areas, available by API and AI-services according to the SaaS model.

More and more Russian companies are transferring work with AI to the cloud. This is due to the fact that training modern large language models and ensuring their performance requires a huge amount of computing resources, as well as attracting expensive specialists. Not every company can afford such an investment. To meet the growing demand of the business, MWS increased the GPU capacity available to customers by 40%. And by the end of 2024, this figure will increase by another 3-4 times. We strive to ensure that as many companies as possible, regardless of their size, have the opportunity to train and introduce their AI models into the business, "said Aleksei Kuznetsov, director of new cloud products at MTS Web Services.

Launch of virtual infrastructure with GPU accelerators at Fedorovsky data center

MTS Web Services (MWS), part of the group, MTS announced the launch virtual of an infrastructure cluster with GPU accelerators DPC in Fedorovsky in. Leningrad Region MTS announced this on August 8, 2024. This segment will speed up and simplify implementation artificial intelligence the Russian by companies. By the end of 2024, there will also data center be a classic cloud segment designed to transfer and store IT business systems in it.

Fedorovsky became the first MWS data center St. Petersburg with GPU cards that are necessary for working with artificial intelligence. The launch of the new segment will allow companies to simplify and speed up training and performance of their own - and AI- ML models, placing them in the MWS cloud. The infrastructure is suitable for working with any AI models in everyone. industries economies

By the end of 2024, the data center will also deploy the classic virtual infrastructure and other popular MWS cloud services. In addition, the company's customers working with artificial intelligence, in addition to the GPU infrastructure, will have access to a platform for the development, training and deployment of ML models, ready-made AI models for different industries and areas, available via API and AI services based on the SaaS model.

More and more Russian companies are transferring work with AI to the cloud. This is due to the fact that training large language models and ensuring their performance requires a huge amount of computing resources, as well as attracting expensive specialists. Not every company can afford such an investment. To meet the growing demand of the business, MWS will quadruple the GPU capacity available to customers by the end of the year. We strive to ensure that as many companies as possible, regardless of their size, have the opportunity to train and introduce their AI models into the business, "said Aleksei Kuznetsov, director of new cloud products at MTS Web Services.

Adding products from Arenadata

MTS, a digital ecosystem, on July 26, 2024 announced the conclusion of a strategic partnership between MTS Web Services (MWS), part of the MTS group, and Arenadata. As part of the contract, MWS launched five services on software products from Arenadata. They will allow the business to deploy databases for management decisions without unnecessary costs based on the analysis of a large amount of information, as well as simplify the training of artificial intelligence models.

The first service was cloudy the database management system () DBMS for storages analyzing large amounts of structured data -. Arenadata DB (ADB) The service is suitable for creating enterprise analytical repositories for financial reporting, client analytics systems marketing , and campaign management. Also, on the basis of the service, you can build enterprise systems: ADB with business analysts any are integrated BI systems that are compatible with JDBC/ODBC. In addition, the service allows you to quickly expand and minimize sandbox zones for pilot projects and testing statistical hypotheses, as well as work with all analytical tools in a single environment.

The second product is an enterprise distribution for distributed scalable data processing - Arenadata Hadoop (ADH). The service is suitable for storing and processing large volumes of structured, semi-structured and unstructured data, creating an infrastructure for Big Data analysis, solving data analysis (DS) and machine learning (ML) problems, as well as continuous collection and analysis of various metrics and logs.

The Arenadata QuickMarts Cluster Column Database (ADQM) will also be available to MWS customers, which allows analytical queries to be performed on large amounts of structured data in real time. The service can be used to quickly generate BI reporting, as well as implement a layer of data marts for user load. ADBs, together with ADQM, are a complete replacement for SAP BW and can be used to create BI systems with more than a thousand users capable of processing requests to hundreds of storefronts.

The fourth and fifth products available at MWS were Arenadata Postgres transactional (OLTP) database (ADPG) and Arenadata Streaming (ADS) data streaming product. Arenadata Postgres is a universal DBMS for working with loads of different profiles (primarily OLTP - Online Transaction Processing). It allows you to work with various amounts of data and supports a wide range of data types (including JSON and user types) with the ability to flexibly program data processing mechanisms. Arenadata Streaming is a scalable fault-tolerant system for real-time streaming, adapted for enterprise use and based on Apache Kafka and NiFi. The product includes a graphical web user interface for managing and monitoring streaming clusters (ADS Control, ADSC), support for advanced security (Arenadata Platform Security, ADPS), as well as a number of additional tools for implementing replication, data quality verification and other functions necessary for product operation.

Russian business is increasingly digitalized. The amount of accumulated data is growing every day. Information about sales, customers, seasonal demand, production processes - all this can become a powerful booster for the growth of the company, if you are able to quickly analyze large amounts of data and make decisions based on data. This can be implemented using specialized services such as a joint product from MWS and Arenadata. The MWS cloud allows a business to quietly accumulate gigantic amounts of information without thinking about storage and investment in physical infrastructure. While data services provide quick access and processing of accumulated information for its subsequent use. Cloud data storage also simplifies artificial intelligence training, as ML models will be trained on gigantic volumes of structured and unstructured data located in one place, said Mikhail Tutaev, MWS Product Director.

Arenadata is actively expanding the presence of its products in the clouds, new confirmation of which was the partnership with MWS. It is important for us that our customers do not become hostages of only one option for building an IT infrastructure, and can work with data in convenient environments for them. At the same time, they receive such significant advantages as flexibility and speed, which is achieved thanks to the ability to quickly receive computing resources without large start-up investments. Also, by choosing Arenadata products in the clouds, customers can make the costs of computing infrastructure variable and more transparent, "said Anton Blizgarev, director of cloud business development at Arenadata.