Ethernet Meets the Filing Cabinet: NAS and SAN in the Early ’90s

This is the fifth blog post in a 12-part series charting the storage journey — decade by decade, technology by technology — showing how the cost-per-GB cliff, how networking advances such as NVMe over TCP enable high-performance data access, and how software innovation got us from washing machines to the frictionless, fabric-native storage we’re building today.

By the early 1990s, storage had outgrown the server room. Faster disks, cheaper compute, and rapidly improving networks were converging, and IT teams finally had the bandwidth to think differently: what if data didn’t have to live inside the machine that used it?

That question gave rise to networked storage—SANs for high-performance block access, NAS for simple file sharing—and quietly redefined how the world worked. Because when storage got cheaper, faster, and more reliable, it didn’t just change IT architecture. It changed the trajectory of modern life.

The Networking Shift: From Islands to Infrastructure

The early 1990s marked the end of network isolation. Token Ring and IPX/SPX were still hanging on, but Ethernet and TCP/IP were taking over—simple, inexpensive, and universal. Fast Ethernet arrived mid-decade, pushing bandwidth to 100 Mbps; Gigabit was right behind it. Suddenly, networks could do more than move email and print jobs—they could move data at a scale no one had seen before.

That convergence of ubiquitous connectivity and standardized protocols made it possible to treat storage as a shared service rather than a peripheral. It’s hard to overstate how transformative that was: it meant data could flow anywhere inside an organization—between users, systems, and even continents—without ever leaving digital form.

SAN: The High-End Foundation of Shared Data

At the enterprise level, the first wave of networked storage came through the Storage Area Network (SAN). Fibre Channel emerged as the connective tissue, marrying SCSI’s trusted command set with the speed and reach of networking. SANs allowed multiple servers to share the same centralized RAID arrays, achieving levels of availability and performance that direct-attached systems couldn’t touch.

It wasn’t cheap—specialized HBAs, optical cabling, and proprietary switches made SANs a luxury for banks, defense labs, and Fortune 500s—but they solved a critical problem: data availability. For the first time, storage could scale and survive hardware failures without interrupting business. That reliability was the foundation for always-on, globally distributed operations.

NAS: Shared Storage for the Rest of Us

SANs proved what was possible; Network Attached Storage (NAS) made it practical. Instead of serving raw blocks, NAS presented shared files over TCP/IP networks using NFS for UNIX and SMB/CIFS for Windows and OS/2. Vendors like NetApp, Auspex, and later EMC Celerra turned NAS into an appliance—plug it into Ethernet, mount it, and go.

Suddenly, everyone could collaborate on the same files without copying them between machines. Engineers shared CAD drawings, marketers shared documents, and universities hosted research datasets accessible across departments. NAS democratized the concept of shared data, delivering it to mid-market IT budgets that could never afford Fibre Channel.

The Expanding Operating System Universe

Operating systems mirrored this transition. Classic UNIX platforms still ruled high-performance computing, but Windows NT 4.0 brought networking and storage services into the mainstream. IBM OS/2 lingered in corporate back offices, and Linux emerged as an open-source UNIX alternative that embraced TCP/IP from day one. Each OS leaned on shared storage—whether NAS or SAN—to manage collaboration, growth, and centralization.

This diversity forced the standardization of protocols like NFS and SMB, effectively teaching the industry to cooperate. It also pushed vendors toward interoperability—a rare word in the 1980s. Storage was no longer an island; it was a network citizen.

Centralization, RAID, and the Birth of Data Management

As data volumes grew, RAID moved from an academic concept to an industry standard. Hardware controllers improved, parity calculations became hardware-accelerated, and enterprise arrays reached capacities measured in hundreds of gigabytes. For the first time, storage could be large, fast, and resilient.

This new reliability created a business revolution: centralized data management. Instead of chasing files across dozens of servers, IT could manage permissions, quotas, replication, and backups from a single point of control. Data became a corporate resource—organized, protected, and continuously available. The role of the storage administrator was born.

The Broader Shift: How Storage Changed the World

The 1990s marked the time when storage stopped being an engineering curiosity and became the backbone of modern civilization. Cheaper, denser, and more resilient disks fueled an explosion in digital creation. Without accessible storage, the world couldn’t have sustained the explosion of the Internet, web servers, or multimedia content. Every website, every email, every photo uploaded, and every transaction logged required space—and for the first time, space was cheap.

If capacity had remained expensive and fragile, innovation would have slowed. There would have been no searchable web, no streaming, no e-commerce, and no cloud computing. What made those revolutions possible wasn’t just connectivity—it was storage availability at scale.

Near-instant access to data, compared to the minutes or hours it once took to retrieve a tape or spin up a job, changed expectations for everything. Businesses could make decisions in real time. Scientists could analyze datasets interactively. Consumers could expect instant information—and that expectation would define the digital era.

The Bridge to Now

The 1990s marked the beginning of the disaggregated, network-centric world we live in today. SANs gave enterprises their first reliable shared infrastructure. NAS appliances made shared storage affordable for everyone else. Together, they turned data into a living system—distributed, resilient, and available on demand.

That same vision drives today’s NVMe/TCP architectures. Ethernet’s universality has replaced Fibre Channel’s specialization; spindles have given way to flash. But the principle hasn’t changed: make storage fast, shared, and always available, and innovation follows.

The 1990s taught us that the cost of storage isn’t just a line item—it’s the price of progress.

Next Up

When Block Went Mainstream: iSCSI and the Ethernet Takeover — how the 2000s proved that “good enough” over TCP/IP could outperform “perfect” over Fibre Channel, paving the way for NVMe/TCP to dominate the modern data center.

To learn more, read more blogs from this series:

About the Writer: