Private blockchain development: platforms, design and trade-offs

A private blockchain is a ledger where only approved organizations run nodes and only approved identities can send transactions. This guide compares the main platforms, explains the design decisions a network needs, and lists the cases where a shared database does the same job with less work.

What a private blockchain is

A private blockchain gives several parties one shared record that no single party can change alone. Each change is signed, ordered by a group of validators and copied to every member's node. Each block holds the hash of the block before it, so a member who edits its own copy breaks that chain of hashes and the other members see the mismatch.

The design pays off when organizations that do not fully trust each other need the same data and none of them should own the database. If one company controls all the data and everyone already trusts it, a normal database is the better tool.

Private, permissioned and consortium networks

People use these three words loosely, and they mean different things.

• A private network in the strict sense has one operator that runs every node, so the network works like an append-only database with signatures.
• A consortium network has several operators, and each member organization runs its own nodes and takes part in governance.
• A permissioned network is any network that checks identity before a node joins or a user transacts. Both private and consortium networks are permissioned.

Most enterprise projects with a clear business case are consortium networks, because the benefit comes from removing the need to trust one operator.

Private blockchain platforms compared

Four platforms cover most private blockchain development today. Their status changed in 2024 and 2025, so the table shows the current state.

Fabric and Besu are projects of LF Decentralized Trust, the Linux Foundation group that took over the Hyperledger projects in 2024. The site has separate guides to Hyperledger development, Corda and Quorum.

Choosing between them

• Besu suits teams that already know Ethereum tools, or that may move assets to a public chain later.
• Fabric suits workflows where different member groups must see different data, because channels split the ledger.
• Corda suits regulated financial agreements, where only the parties to a deal should receive its data.

Team skills and the preferences of the member organizations often decide the choice more than features do.

Private blockchain development: designing the network

Governance before code

The questions that decide whether a consortium works are not technical, and the members must agree on these points:

• who may join, and who approves new members
• who pays for nodes and shared services
• how smart contract upgrades get approved
• what happens to data and nodes when a member leaves
• which country's law governs disputes

Write the answers into a consortium agreement before you choose a platform.

Consensus

Permissioned networks reach agreement by voting among known validators, not by mining. Besu's QBFT is Byzantine fault tolerant: it keeps working while fewer than one third of validators are faulty or dishonest, so a network of four validators survives one bad validator.

Fabric's Raft ordering tolerates crashed nodes but not dishonest ones, which is acceptable when one trusted party runs all the orderers. Fabric v3.0 added a SmartBFT ordering service for networks where up to a third of orderer operators may not be trusted, and channels must enable the V3_0 capability to use it.

Privacy

Most businesses cannot accept a ledger where every node sees every transaction. Each platform offers a different answer:

• Fabric has channels and private data collections. A private data collection shares data only with named organizations and puts a hash on the shared ledger.
• Corda sends each transaction only to the parties that need it.
• Besu privacy groups relied on Tessera, and the Besu maintainers deprecated that feature in 2024, so new Besu networks should handle privacy at the application layer.
• Any platform can store only hashes on-chain and keep the documents in each party's own systems. This pattern is simple and works on every platform.

Personal data

Ledger records cannot be deleted, and data protection law gives people the right to have their data erased. In April 2025 the European Data Protection Board published draft guidelines on blockchain and personal data, which tell controllers to avoid storing personal data on-chain. Keep personal data off-chain and store only a hash or a reference on the ledger.

Identity and keys

Permissioned chains rely on a public key infrastructure, and each organization usually runs its own certificate authority. Signing keys belong in hardware security modules, and the design also needs a plan for key rotation and for revoking a certificate when an employee leaves.

Integration and operations

The ledger itself is often less than a quarter of the project effort. Most of the work goes into the systems around it:

• connecting ERP systems, document stores and identity providers
• writing event listeners that push ledger updates into existing workflows
• building dashboards for staff who never see the chain
• running nodes with monitoring, backups and upgrade procedures across organizations with different IT stacks

When a database is better

• One organization controls the data and everyone already trusts it.
• Records must be corrected or deleted freely.
• Throughput and latency targets are strict, and there is no multi-party trust problem.
• The members will not run their own nodes. A consortium where one vendor hosts every node gives no more protection than that vendor's database.

Before a consortium commits to production, a short blockchain proof of concept can test whether the multi-party benefit is real.

Timelines

A pilot with two or three members, one workflow and one platform usually takes three to six months, and governance and integration take most of that time. A production rollout across many organizations often takes a year or more. If a vendor promises a production consortium in a few weeks, ask which members have signed the governance agreement.

The guide to enterprise blockchain solutions covers other enterprise options.

Frequently asked questions