About the Course

A theoretical course covering various aspects of data management, including computing systems, virtualization, data storage, and network technologies in modern data centers.

Contains information on components and types of computing systems, data storage, network technologies, as well as software for data management in data centers.

Target Audience

• IT Architects
• IT Managers
• Storage Pre-sales

Advantages

The course is designed for storage architects and technical specialists responsible for planning, selection, configuration, and development of data storage systems. Even for those confident in their knowledge of the covered topics, the course will be useful for its practical exercises – after all, it's not always possible to test skills in a production network. If even the course's practical exercises are insufficient, you can take a separate training focused solely on performing laboratory work to improve your skills.

Day 1

File, Block, and Unified Storage Systems

Operating Principles

• File Systems (NAS): organize data as a hierarchy of files and directories, using CIFS/SMB, NFS protocols. Data access is via file names.
• Block Systems (SAN): provide data access at the block level (LUN). Computing systems create a file system on top of block volumes. Use iSCSI, FC, FCoE protocols.
• Unified Systems: combine block, file, and object access in a single infrastructure. Support multiple interfaces (iSCSI, FC, NFS, CIFS, REST) via a single controller.

Key Components

• File System (NAS): file server with a file system; network interfaces (Ethernet); data storage (typically HDD/SSD).
• Block System (SAN): controllers (with components: frontend, cache, backend); block storage devices (HDD, SSD); storage area network (SAN) based on FC, iSCSI; caching mechanisms (write‑through, write‑back).
• Unified System: single controller supporting different access types; integrated interfaces (block, file, object); common pooled storage.

Main Features

• File System: user-friendly, easy file management, scaling limitations (number of files/directories).
• Block System: high performance for transactional loads, efficient storage management, complexity for unstructured data.
• Unified System: reduced costs through unified infrastructure, centralized management, flexibility for different applications.

Data Protection

• RAID
• Snapshots and Clones
• Thin Provisioning
• Deduplication and Compression
• Replication (synchronous/asynchronous)
• Antivirus Protection

Scaling

• File System: limited by directory hierarchy and metadata performance.
• Block System: scale‑up (expanding capacity of a single array) or scale‑out (controller clustering).
• Unified System: linear scaling by adding nodes, support for distributed storage pools.
Day 2

Object Storage Systems

Operating Principles

• Uses flat address space.
• Data is stored as objects, containing: user data; metadata (size, owner, date, attributes); unique object ID (generated via hash function from content).

Key Components

• Nodes (controllers) with two services: metadata service (generates ID, manages namespace); storage service (manages disks, places objects).
• Internal network for node communication.
• Physical storage based on inexpensive high-density disks.
• Global namespace for unified object access.

Main Features

• Linear scalability (scale‑out) – adding nodes increases capacity and performance.
• Multi-tenancy – secure data separation between applications/clients.
• Metadata-based management – automation of storage, protection, data movement policies.
• Flexible access methods: S3, REST/SOAP, CIFS, NFS, HDFS, HTTP, GRPS.
• Automation: self-configuration, self-healing, minimizing manual management.

Data Protection

• Replication – creating exact copies of objects.
• Erasure coding (error correction coding) – optimal redundancy for disk failures.
• Data encryption before sending to the cloud (via gateways).
• Deduplication and compression for capacity savings.

Scaling

• Support for petabyte and exabyte data volumes.
Day 3

Security, Performance, Use Cases, and Incident Management

Security and Access Management

• Authentication and Authorization Mechanisms: use of protocols and standards for data access control; access rights separation based on roles and security policies for different storage types (file, block, object, unified).
• Data Encryption: at transmission level (e.g., TLS/SSL in REST API for object storage); at storage level (encryption of object content, block LUNs, file systems); encryption of SAN (FC, iSCSI) and NAS network traffic.
• Metadata Access Control: in object systems – management via object metadata; in file systems – via ACL (Access Control Lists); in block systems – at LUN and network zoning levels (FC zoning).
• Logging and Auditing: tracking data operations (creation, modification, deletion); monitoring access to critical objects and data blocks.

Performance Tuning

• Cache Optimization: configuring caching algorithms in block systems (write‑through vs. write‑back); managing cache size and eviction policy; using SSD cache to accelerate access to "hot" data.
• Load Balancing: distributing I/O between controllers in SAN; balancing requests between nodes in scale‑out systems (object, unified); aggregating network channels to increase throughput.
• Storage Tiering: automatic data movement between storage tiers (SSD → HDD) based on activity; cache tiering; object placement policies based on metadata.
• Deduplication and Compression: reducing data volume in storage (especially effective for backups and archives); assessing CPU overhead for processing compressed/deduplicated data.
• RAID Parameter Tuning: choosing RAID level (1, 5, 6, 10) for the type of load (transactional, analytical); considering the impact of RAID on write performance.

Use Cases and Best Practices for Each System Type

• File Systems (NAS): shared file access in corporate environments; storing documents, media content, archives; integration with office applications and document management systems. Best practices: regular metadata optimization, monitoring directory nesting depth.
• Block Systems (SAN): databases (OLTP, DBMS); virtualized environments (VMware, Hyper‑V); high-performance computing (HPC). Best practices: allocating separate storage pools for different load types, using SSDs for "hot" data.
• Object Systems (OSD): cloud services (SaaS, BaaS, AaaS); storing large volumes of unstructured data (logs, media, IoT data); big data analytics (integration with HDFS); distributed applications with global access. Best practices: applying erasure coding for space savings, configuring object lifecycle policies.
• Unified Systems: hybrid and multi-cloud environments; consolidation of heterogeneous workloads (databases + files + objects); Best practices: centralized storage policy management, automating tiering between levels.