There's a question that has been quite thought-provoking in the industry—given the variety of storage solutions, why spend years developing your own system? Let's take a look at Walrus's approach.

**The Old Problem of Data Accumulation**

The traditional blockchain design logic is this: to achieve decentralization, every node must run a full copy of the data. Sounds straightforward, but the real trouble begins— as on-chain data grows, the hard drive capacity of ordinary users becomes a bottleneck. When hard drives fill up and nodes can't run anymore, what happens? The network becomes increasingly centralized.

This industry has been stuck in this dead end: for security, full redundancy is necessary (like Bitcoin does), but for scalability, security must be weakened. A dilemma.

**Walrus's Trick—Splitting It Up**

The core idea is actually quite simple: not everything needs to be handled by a single system, so why not split it up?

Walrus divides the system into two layers:

One side is the **Control Layer**, built directly on Sui. Why not build a chain of your own? Because Sui already handles ultra-high concurrency and instant confirmation, so there's no need to reinvent the wheel. What does this layer mainly do? — Node staking management, storage capacity transaction settlement, file metadata indexing, system parameter adjustments. Sui's high performance naturally provides this control layer with strong throughput and determinism.

The other side is the **Storage Layer**, which is the actual place where data is stored. These two layers operate separately, without interference.

By splitting like this, the issues of storage expansion and consensus complexity can be addressed separately. No need to cram everything onto a single chain. It may seem like a small change, but it actually transforms the entire approach.