Storage Virtualization

Understanding storage virtualization techniques including SAN, NAS, software-defined storage, and how cloud providers deliver scalable storage services.

Storage virtualization abstracts physical storage devices into logical storage pools, presenting a unified view of storage resources regardless of their physical location or type. It enables efficient storage management, flexible allocation, and advanced features like thin provisioning and snapshots.

Storage Virtualization Architecture

│ │ VM sees	/dev/sda (100GB disk) │ │
│ │ App sees	Mounted filesystem │ │
│ │ Features	Thin provisioning, Snapshots, │ │
│ │ Total Physical	24TB │ │
│ │ Presented to VMs	Up to 100TB (thin provisioned) │ │

Types of Storage Virtualization

Block-Level Virtualization (SAN)

bash example

# Linux LVM - Block storage virtualization
# Create physical volumes
sudo pvcreate /dev/sdb /dev/sdc /dev/sdd

# Create volume group (storage pool)
sudo vgcreate storage-pool /dev/sdb /dev/sdc /dev/sdd

# Create logical volumes (virtual disks)
sudo lvcreate -n web-data -L 500G storage-pool
sudo lvcreate -n db-data -L 1T storage-pool

# Thin provisioning (allocate more than physical capacity)
sudo lvcreate --thin -n thin-pool -L 2T storage-pool
sudo lvcreate --thin -n vm-disk1 -V 500G --thinpool thin-pool storage-pool
sudo lvcreate --thin -n vm-disk2 -V 500G --thinpool thin-pool storage-pool

# Snapshots
sudo lvcreate --snapshot -n db-snapshot -L 50G storage-pool/db-data

File-Level Virtualization (NAS)

bash example

# NFS Server setup (Network Attached Storage)
sudo apt install nfs-kernel-server
sudo mkdir -p /exports/shared-storage

# Configure NFS exports
echo "/exports/shared-storage 10.0.0.0/24(rw,sync,no_subtree_check)" | \
  sudo tee -a /etc/exports
sudo exportfs -a
sudo systemctl restart nfs-kernel-server

# Mount on client VM
sudo mount -t nfs server:/exports/shared-storage /mnt/shared

Software-Defined Storage (SDS)

Example

# Ceph distributed storage cluster # ceph.conf [global] mon_initial_members = mon1, mon2, mon3 mon_host = 10.0.0.1, 10.0.0.2, 10.0.0.3 auth_cluster_required = cephx auth_service_required = cephx auth_client_required = cephx osd_pool_default_size = 3 osd_pool_default_min_size = 2 # Create a storage pool ceph osd pool create vm-storage 128 128 # Create a block device for a VM rbd create --size 100G vm-storage/vm-disk-1 # Map the block device rbd map vm-storage/vm-disk-1

Cloud Storage Virtualization

Service	Type	Provider	Use Case
EBS	Block	AWS	VM disks, databases
S3	Object	AWS	Files, backups, data lakes
EFS	File	AWS	Shared filesystem for containers
Azure Managed Disks	Block	Azure	VM storage
Persistent Disk	Block	GCP	Compute Engine VMs
Cloud Storage	Object	GCP	Unstructured data

bash example

# AWS EBS - Create and attach virtual block storage
aws ec2 create-volume \
  --availability-zone us-east-1a \
  --size 1000 \
  --volume-type gp3 \
  --iops 16000 \
  --throughput 1000 \
  --encrypted

# Storage features in cloud
# Snapshots (point-in-time backup)
aws ec2 create-snapshot --volume-id vol-123456 --description "Daily backup"

# Cross-region replication
aws s3api put-bucket-replication --bucket my-bucket \
  --replication-configuration file://replication.json

Key Features of Storage Virtualization

Thin Provisioning: Allocate more storage than physically available
Snapshots: Point-in-time copies without full data duplication
Deduplication: Eliminate duplicate data blocks
Compression: Reduce storage space used
Tiering: Automatically move data between fast (SSD) and slow (HDD) storage
Replication: Mirror data for disaster recovery

Interview Questions

What is storage virtualization and why is it important?

Storage virtualization abstracts physical storage into logical pools, enabling flexible allocation, thin provisioning, snapshots, and unified management. It's important because it improves utilization (from ~30% to 70%+), simplifies management, and enables advanced data services.

Explain thin provisioning with an example.

Thin provisioning allocates virtual storage that exceeds actual physical capacity. Example: With 10TB physical storage, you can provision 10 VMs with 5TB virtual disks each (50TB total). Actual space is consumed only as data is written. This works because most VMs don't use their full allocation.

What is the difference between SAN and NAS?

SAN provides block-level storage over a dedicated network (Fibre Channel or iSCSI) — VMs see raw disks. NAS provides file-level storage over standard networks (NFS, SMB) — VMs see shared directories. SAN is for high-performance databases; NAS is for shared file access.

How does Ceph implement software-defined storage?

Ceph distributes data across commodity servers using CRUSH algorithm for placement. It provides block (RBD), file (CephFS), and object (RADOS) storage from a single cluster. Data is replicated across nodes for reliability, and the cluster self-heals when nodes fail.

How do cloud providers implement storage virtualization at scale?

Through distributed storage systems that replicate data across multiple physical devices and availability zones. For example, AWS EBS replicates within an AZ, S3 replicates across AZs. Storage is provisioned via API, scaled elastically, and billed per-GB per-month.