Computer Memory and Storage

IBM invented the first hard disk drive (HDD) in 1956 as part of the RAMAC (Random Access Method of Accounting and Control) system. The RAMAC 350 stored 5 MB of data on 50 24-inch platters and was the size of two refrigerators. IBM's innovation made random access to large amounts of data possible, revolutionizing data storage. Apple, Microsoft, and Western Digital entered the storage market much later.

Intel 1103, introduced in 1970, was the world's first commercially available dynamic RAM (DRAM) chip. It stored 1024 bits (1 Kb) and replaced magnetic core memory, enabling smaller, cheaper computers. SDRAM, DDR, and Intel 504 are all later technologies or nonsensical options. The 1103 established Intel as a memory chip leader before they shifted to microprocessors.

A synthetic full backup consolidates the last full backup with all subsequent incremental backups to create a new full backup file without reading all data from the source. This is more efficient than a traditional full backup while providing the same recovery benefits. Differential backups would not be used in this process.

In Server-free backup architecture, data moves directly from storage to backup device via the SAN, completely bypassing the server's CPU and memory. The server only initiates the backup operation but is not involved in the actual data transfer, freeing it for other tasks. LAN-free backup still uses the server's resources.

Replication in backup/storage contexts typically involves writing data to the primary storage first, then copying it to secondary storage. This distinguishes it from mirroring (option C), which typically implies simultaneous writes to multiple locations. Option A describes synchronous mirroring, not replication.

A snapshot is a point-in-time, read-only copy of data that captures the state of a storage system at a specific moment. Unlike a full backup or data copy, a snapshot uses copy-on-write mechanisms and metadata pointers to be space-efficient, allowing quick recovery to that exact state.

Hard disk drives store data using magnetic technology - tiny magnetic domains on spinning platters represent binary data. Laser and light are used in optical storage (CDs, DVDs), while chemical storage is not used in consumer HDDs.

NAS (Network Attached Storage) is specifically designed for file-level storage and sharing over a network. It provides file-based data access to multiple clients, with built-in file systems and protocols like NFS and SMB. SAN offers block-level storage better suited for databases, while DAS directly attaches to a single server without network sharing capabilities.

RAID 5 uses distributed parity across all disks and requires a minimum of 3 disks to function. With 2 disks, you can only create RAID 0 (striping) or RAID 1 (mirroring). RAID 5 needs at least 3 disks to distribute parity blocks while maintaining data redundancy - it can survive a single disk failure.

RAID 1 (mirroring) provides high availability by duplicating data across two or more disks. If one disk fails, the system continues operating seamlessly from the mirror. RAID 0 offers no redundancy (striping only), RAID 2 is obsolete and rarely used in practice. Mirroring is the simplest HA solution.

SAN (Storage Area Network) provides the fastest data movement between server and storage because it uses dedicated, high-speed network infrastructure like Fibre Channel (up to 32Gbps or more). DAS is limited by server bus bandwidth, while NAS has protocol overhead and shared network contention. SAN's dedicated path and block-level access optimize for speed.

Tape libraries connect to SANs because tape drives require Fibre Channel connectivity for backup and archival workflows in enterprise environments. NAS uses file-level protocols (NFS/SMB) inappropriate for tape, while DAS lacks the network architecture. SAN provides the necessary fabric for integrating tape into the storage infrastructure.

SAN is the preferred storage for databases because it provides high-speed, low-latency, block-level access that databases require. NAS adds file-layer overhead and latency, while DAS lacks scalability and sharing capabilities. Database workloads need consistent, fast I/O with multiple server access - SAN's dedicated infrastructure delivers this optimally.