Choosing the right high availability architecture is critical for balancing cost, complexity, and reliability. The two main approaches — shared nothing and shared disk — differ fundamentally in how they handle data redundancy and failover. A shared nothing architecture like SafeKit replicates data between independent servers with their own local storage, while a shared disk architecture relies on a centralized storage bay (SAN) accessed by multiple servers. The following table compares these two approaches across key criteria.
| Criteria |
Shared Nothing Architecture |
Shared Disk Architecture |
| Product |
SafeKit on Windows and Linux |
Clustering toolkit for shared disk |
| Extra hardware |
No — uses internal disks of servers |
Yes — extra cost with a shared bay of disks (SAN) |
| Application data organization |
Zero impact on application data organization. Just define directories to replicate in real-time. Even directories inside the system disk can be replicated. |
Requires special configuration of the application to put its data on a shared disk. Data on the system disk cannot be recovered. |
| Complexity of deployment |
No — install a software on 2 servers |
Yes — requires specific IT skills to configure OS and shared disk |
| Failover |
Just restart the application on the second server. |
Switch the shared disk. Remount the file system. Run the recovery procedure on the file system. Then restart the application. |
| Disaster recovery |
Just place the 2 servers in 2 remote sites connected by an extended LAN. |
Extra cost with a second bay of disks. Requires specific IT skills to configure mirroring of bays across a SAN. |
| Quorum and split brain |
Application runs on a single server after network isolation (split brain). Data coherency is guaranteed. No need for a third machine, quorum disk, or special heartbeat line. More on heartbeat, failover and quorum |
Requires a special quorum disk or a third quorum server to avoid data corruption on split brain. |
| Best suited for |
Software editors wanting to add a simple HA option to their application |
Enterprises with IT skills in clustering and large database applications |
In summary, the shared nothing architecture, as implemented by SafeKit, wins on simplicity, cost, and deployment speed. It requires no extra hardware, no specialized IT skills, and no changes to the application's data organization. Failover is straightforward, disaster recovery is built-in, and split brain is handled without additional infrastructure.
The shared disk architecture remains relevant for large enterprises with existing SAN infrastructure and dedicated IT teams managing very large databases. However, for most application high availability needs — especially for software editors, OT/edge environments, and distributed sites — shared nothing with SafeKit is the simpler, more cost-effective choice.
