New storage capabilities, local CDN, the future of routing in Kubernetes® and other Yandex Cloud updates

In early November, the flagship Yandex Cloud conference, Yandex Neuro Scale Kazakhstan, took place in Almaty, focusing on the application of AI and cloud solutions for business growth.

The Infra & DevTools track became one of the key sessions for specialists interested in tools for optimizing infrastructure, data storage, and traffic routing. The main insights are summarized in our article.

Storage Evolution: New S3 Capabilities and Cloud Backup

Alexander Kirov, Technical Manager at Yandex Cloud, responsible for the development of data storage infrastructure services, discussed the main updates to the S3 object storage and the cloud backup service, which has become available to businesses in Kazakhstan.

In Object Storage, authentication via static keys and STS tokens has been joined by authentication using IAM tokens. There is no longer a need to create a service account and obtain static keys for it. An IAM token is a secure authentication method for automated scenarios, CI/CD pipelines, and infrastructure deployment. The token can be used via the S3 API, Terraform, and CLI.

Also this year, Yandex Cloud implemented full-featured support for working with buckets and objects within them via the YC CLI utility. Now, for many scenarios, this is sufficient. For example, you can create a bucket, upload a virtual machine image to it, and launch a virtual machine from that image.

Alexander emphasized that when adding features, Yandex Cloud focuses on the most in-demand scenarios. For a popular task—evaluating the efficiency of bucket usage—the company will soon introduce the S3 Inventory tool. It will allow generating a dump of bucket contents at a specific point in time in the form of a table. It will include all object metadata: size, creation date, MD5 checksum, storage class, versioning data, ACL, and more.

Tabular export of metadata for all objects opens up wide opportunities for analytics, allows finding duplicate files, simplifies data synchronization between different storage systems and services, improves audit convenience, as well as backup and versioned data recovery scenarios.

Alexander shared insights on rapidly growing use cases for object storage, highlighting three main ones:

1. S3 together with modern databases such as ClickHouse® and GreenPlum®: rarely used data can be offloaded to a cheaper cold storage.

2. S3 as a scalable universal layer for machine learning and data processing scenarios: fast data access, ease of data delivery, and low storage costs.

3. S3 as the primary layer for secondary data—archives and backups of data that are in active use.

This year, Cloud Backup also became available in Kazakhstan—a specialized service for backing up resources within Yandex Cloud. It allows creating consistent backups of virtual machines, restoring files and entire machines, managing backup schedules, and protecting resources during updates.

Within Cloud Backup, a backup and recovery agent is installed in each virtual machine. The agent allows creating copies of virtual machine disks according to a configured policy or on demand, while synchronizing its backup operations with the operating system and file system.

Data can be restored to the original or a new virtual machine in the cloud. Individual file restoration is also available: this is useful, for example, when you need to quickly restore only configuration files rather than the entire system.

Yandex Cloud CDN: How It Works Under the Hood

Evgeny Zaitsev, Senior Technical Project Manager at Yandex Cloud, explained how the CDN service used by Yandex is structured.

CDN in Kazakhstan provides a fast and efficient way to deliver content to end users:

— reduces latency and accelerates application response
— optimizes traffic costs
— increases service stability and reliability
— easily integrates with existing infrastructure.

Yandex operates six local CDN locations within Kazakhstan and more than 80 points of presence outside the country. This ensures high speed and scalability for global delivery of data from various company services.

Currently, 80% of Yandex's traffic in Kazakhstan is distributed via CDN. The service supports a throughput of over 150 Gbps, ensuring comfortable access to media content and other applications without network overload.

The Cloud CDN service on the Yandex Cloud platform will become available to users in Kazakhstan in the first half of 2026.

The Future of Routing in Kubernetes®: Yandex Cloud Gwin

Nursultan Kalniyazov, Solutions Architect for Kubernetes® Services at Yandex Cloud, presented the new Yandex Cloud Gwin controller for managing traffic in Kubernetes® and shared plans for product development.

The Gwin controller supports two specifications simultaneously: Ingress and Gateway API. Thus, it retains all the advantages of ALB: flexible routing and automatic scaling. Now developers can manage external and internal traffic through a single tool without losing compatibility with familiar APIs.

Key advantages of Yandex Cloud Gwin:
— a single controller for various routing scenarios
— flexible configuration of unique ALB features
— scaling under load.

In the Infra & DevTools track, Yandex Cloud experts demonstrated a wide range of tools for data storage, routing, logging, backup, and infrastructure scaling. The Yandex Neuro Scale Kazakhstan conference clearly showed that technology in Kazakhstan is actively developing, adapting to growing loads and business needs.