A blockchain is an append-only ledger where records are grouped into blocks and linked together using cryptographic hashes. Bitcoin processes roughly 7 transactions per second across its entire network — compared to Visa’s 65,000 TPS capacity — which illustrates the fundamental throughput cost of decentralized consensus and why blockchain is only justified when mutual distrust between parties makes a central authority impossible.

The hash link makes tampering detectable: changing a past block changes its hash, which breaks the chain unless all subsequent blocks are recomputed (and the network’s consensus rules are satisfied).

Example

Toy example: each block commits to the previous block hash.

using System;
using System.Security.Cryptography;
using System.Text;
 
static string Sha256Hex(string s)
{
    var bytes = Encoding.UTF8.GetBytes(s);
    var hash = SHA256.HashData(bytes);
    return Convert.ToHexString(hash).ToLowerInvariant();
}
 
var genesisPrev = new string('0', 64);
var genesisData = "genesis";
var genesisHash = Sha256Hex(genesisPrev + genesisData);
 
var block2Prev = genesisHash;
var block2Data = "tx: alice -> bob (10)";
var block2Hash = Sha256Hex(block2Prev + block2Data);

Pitfalls

Using Blockchain When a Database Suffices

What goes wrong: teams adopt blockchain for internal systems where all parties trust a central authority, gaining none of the decentralization benefits while paying the full cost in throughput, complexity, and compliance risk. An Australian government agency spent AUD $8.5M on a blockchain-based supply chain system that was eventually replaced by a PostgreSQL database with audit logging — all participants were government departments that already trusted a central authority, so the consensus mechanism added latency and complexity with zero benefit.

Why it happens: blockchain is associated with innovation and security, making it attractive even when the problem doesn’t require it.

Mitigation: blockchain is justified only when you need a shared ledger across mutually distrusting parties with no central authority. If you control all the nodes, a traditional database with audit logging is simpler, faster, and easier to comply with GDPR.

GDPR Conflict with Immutability

What goes wrong: storing personal data on a blockchain makes it impossible to fulfill GDPR’s right to erasure (Article 17). Once written, the data cannot be deleted without breaking the chain.

Mitigation: never store personal data directly on a blockchain. Store a hash or reference; keep the actual data in a mutable off-chain store that can be deleted.

Tradeoffs

Consensus Mechanisms

MechanismThroughputEnergyDecentralizationUse when
Proof-of-Work (PoW)~7 TPS (Bitcoin)Very high (ASIC mining)HighMaximum censorship resistance; energy cost is acceptable
Proof-of-Stake (PoS)~15–30 TPS (Ethereum)LowHighEnergy efficiency matters; validators stake tokens as collateral
Proof-of-Authority (PoA)Thousands TPSMinimalLow (known validators)Private/consortium chains where validators are trusted entities

Decision rule: PoW for maximum trustlessness (public cryptocurrency). PoS for public chains where energy matters. PoA for enterprise/private chains where you know and trust all validators — but at that point, ask whether a traditional database with audit logging is simpler.

Public vs Private Chains

TypeParticipantsThroughputImmutabilityUse when
Public (Bitcoin, Ethereum)AnyoneLow (7–30 TPS)AbsoluteTrustless, permissionless ledger across unknown parties
Private/Consortium (Hyperledger)Known entitiesHigh (thousands TPS)ConfigurableEnterprise use with known participants; still need shared ledger
Traditional DB + audit logInternalVery highSoft (admin can edit)All parties trust a central authority

Decision rule: if all parties trust a central authority, use a traditional database with append-only audit logging. Blockchain adds value only when you need a shared ledger across mutually distrusting parties with no central authority.

Questions

Limitations for Enterprise Use

  • Throughput: Public blockchains (Bitcoin: ~7 TPS, Ethereum: ~15 TPS) are orders of magnitude slower than traditional databases (thousands of TPS). Private blockchains are faster but lose decentralization benefits.
  • Immutability is a liability: GDPR’s right to erasure conflicts with blockchain’s append-only nature. Storing personal data on a blockchain creates compliance problems.
  • Consensus overhead: Proof-of-Work wastes energy. Proof-of-Stake is more efficient but adds validator complexity.
  • When to use: Blockchain is justified when you need a shared ledger across mutually distrusting parties with no central authority. For most enterprise use cases, a traditional database with audit logging is simpler and faster.

References