Course Syllabus: 18 Modules
From Linux kernel I/O mechanics and RAID to Ceph, Cloud, and Kubernetes storage engineering.
Intro to Data Storage
Data vs information, storage hierarchy, capacity, IOPS, latency, and Linux benchmark tools.
How Storage Actually Works
App to disk I/O path, blocks, sectors, 4K alignment, RMW penalty, and page cache profiling.
Storage Media (HDD, SSD, NVMe)
Platters and seek times vs NAND flash physics, FTL wear leveling, garbage collection, and TRIM.
Interfaces & Protocols
SATA, SAS, Fibre Channel SAN, iSCSI, and hands-on NFS server configuration.
Filesystems Deep-Dive
Inodes, directory trees, ext4 vs XFS vs Btrfs vs ZFS, journaling, and mount tuning.
Block Storage, LVM & RAID
Linux device mapper, LVM PV/VG/LV, RAID 0/1/5/6/10 write penalty, and disk failure simulation.
Object Storage Internals
Block vs File vs Object, S3 architecture, versioning, multipart uploads, and local MinIO lab.
Distributed Storage
Replication models, quorum calculus (R+W>N), sharding, and Reed-Solomon Erasure Coding.
Ceph in Practice
RADOS, MON, OSD BlueStore, CRUSH map determinism, RBD block devices, and cluster recovery.
Databases & Storage
Slotted pages, B-Tree vs LSM, Write Amplification Factor, ARIES WAL crash recovery.
Performance & fio
Workload characterization, Little’s Law, queue depth tuning, and enterprise fio benchmark suites.
Data Protection & DR
Backup vs replication, CoW vs RoW snapshots, RPO & RTO formulas, and 3-2-1 backup strategies.
Cloud Storage (AWS/Azure/GCP)
EBS gp3/io2 vs Managed Disks vs GCS, object lifecycle tiering, and cloud invoice optimization.
Kubernetes Storage & CSI
PV, PVC, StorageClasses, Container Storage Interface plugins, and StatefulSet database lab.
Storage Security & Compliance
LUKS block encryption, TLS in-transit, POSIX ACLs, IAM, and NIST SP 800-88 crypto-shredding.
Architecture & System Design
6-factor storage tradeoffs, capacity planning, and reference patterns for FinTech & streaming.
Troubleshooting Runbook
Diagnostic playbooks: Inode exhaustion, latency spikes, degraded RAID, and slow Ceph OSDs.
Final Capstone Project
Design a 100 TB multi-region storage platform: 30% growth, 99.99% SLA, hybrid DB + object.
Mechanical Sympathy in Real Time
Explore the physical latencies separating registers from persistent flash and spinning disks.
Deep-Dive Architectural Pillars
Specialized technical reference encyclopedia across 7 core domains.
1. Hardware & Physical Layer
Memory hierarchies, NAND Flash cell physics (SLC/TLC/QLC), FTL wear leveling, NVMe over PCIe, and mechanical disk seeking.
2. Kernel & OS Subsystem
How the Linux kernel abstracts block devices: VFS dentries/inodes, dirty page writeback, io_uring ring buffers, and O_DIRECT.
3. Storage Engines & Data Structures
The computational primitives organizing data on disk: B+ Tree slotted pages, LSM SSTables, Bloom filters, and segmented logs.
4. Database Storage Architectures
Transactional durability and analytical engines: Row vs Columnar (OLTP vs OLAP), WAL & ARIES recovery, and MVCC snapshot isolation.
5. Distributed Storage & Consensus
Scaling persistence across fault-prone networks: Consistent hashing rings, tunable quorums (R+W>N), Raft state machines, and S3.
6. Formats & Compression
Binary encoding efficiency: Apache Parquet Dremel shredding, Apache Arrow in-memory IPC, and Zstandard / LZ4 compression.
7. Caching & Memory Management
Algorithms and topologies for high-speed caching: ARC, W-TinyLFU, Cache-Aside vs Write-Behind, and cache stampede solutions.
Engine Decision Engine
Evaluate the RUM conjecture trade-offs for your specific workload.