“The Zilliqa Technical Whitepaper” is the 2017 paper by Zilliqa Research (co-founders Loi Luu, Prateek Saxena, and others from the National University of Singapore) describing the first public blockchain to implement transaction and network sharding in a production mainnet. Unlike Ethereum’s then-theoretical sharding roadmap, Zilliqa demonstrated sharding with real throughput results on a live testnet before any other major chain.

The core claim: by dividing the network into smaller committees that process transactions in parallel, Zilliqa’s throughput scales linearly with the number of nodes — doubling the network size doubles the TPS.

> Whitepaper: Available at zilliqa.com/wp/whitepaper.pdf.

---

Publication and Context

The whitepaper was published in 2017 during the height of Ethereum’s scaling debate. At the time, Ethereum could manage ~15 TPS; Zilliqa’s testnet demonstrated 1,389 TPS with 1,800 nodes — a concrete demonstration that sharding worked in practice.

Zilliqa mainnet launched in January 2019, making it the first (and for some time, only) production blockchain using network + transaction sharding with BFT consensus.

---

Sharding Architecture

Zilliqa’s sharding divides both the network and the transaction set:

Network sharding:

• The full set of miners is divided into shards (committees) of ~600 nodes each
• A Directory Service (DS) Committee of ~800 nodes manages shard formation and global coordination
• Each new DS Committee is elected by PoW (proving membership), then operates via PBFT for all subsequent steps — PoW is used only for Sybil resistance and committee membership, not for block production

Transaction sharding:

• Transactions are assigned to shards by the sharded address space: transactions touching accounts in a given address range go to the corresponding shard
• Each shard processes its transaction subset in parallel and produces a microblock
• Microblocks from all shards are combined by the DS Committee into a final block

Linear scaling claim:

If 600 nodes/shard process T TPS per shard, then N shards process N×T TPS total. As the network grows, new shards form, and throughput scales.

---

Consensus: EC-Schnorr PBFT

Each shard (committee) and the DS Committee run Practical Byzantine Fault Tolerant (PBFT) consensus with EC-Schnorr multisignatures for efficiency:

Standard PBFT produces O(n²) messages. Zilliqa uses multisignature aggregation: validators sign with Schnorr signatures over elliptic curve keys, and a designated leader aggregates 2/3+ signatures into a compact multisignature — reducing broadcast cost to O(n).

Finality: PBFT provides immediate finality — no probabilistic confirmation, no chain reorganization after a block is committed.

---

Scilla Smart Contract Language

Zilliqa introduced Scilla (Safe-by-Design Intermediate-Level LAnguage), a formally verified smart contract language:

• Strict separation of computation and communication: contracts alternate between computation phases and message-passing phases in a structured way
• Eliminates whole classes of reentrancy attacks by design (Scilla contracts cannot call external contracts during computation phases)
• Supports formal verification via model checking

Scilla was designed to prevent the kinds of bugs that caused the DAO hack and Parity multisig vulnerabilities. However, its restrictiveness made it harder to write complex DeFi contracts, and many Zilliqa applications struggled to replicate Solidity’s composability patterns.

---

Sections of the Whitepaper

Section	Content
1. Introduction	Motivation; scalability trilemma context
2. Design Overview	High-level sharding architecture
3. Network Sharding	PoW-based shard formation; DS Committee role
4. Consensus	EC-Schnorr PBFT protocol specification
5. Transaction Processing	Cross-shard transaction handling; microblock construction
6. Security Analysis	Attack models; sybil resistance; DS Committee attack
7. Evaluation	Testnet results: 1,389 TPS at 1,800 nodes

---

Reality Check

Zilliqa demonstrated production sharding years before competitors, which was a genuine achievement. In practice:

• Throughput in production: Mainnet never reached the 1,389 TPS testnet benchmark. Real-world throughput stabilizes at 100–400 TPS due to DS Committee bottlenecks and cross-shard coordination overhead.
• Smart contract fragmentation: Scilla’s safety properties came at the cost of composability. Many DeFi developers avoided Zilliqa due to Scilla’s complexity, slowing DeFi ecosystem growth.
• EVM compatibility added (2022): Zilliqa later added an EVM-compatible layer alongside Scilla to attract Solidity developers — but this architectural decision (two VMs, one chain) created developer confusion.
• Market share: Despite first-mover advantage in sharding, Zilliqa never captured significant DeFi or NFT market share compared to Ethereum L2s, Solana, or Avalanche.

---

Legacy

Zilliqa’s 2017 testnet results were the clearest early evidence that blockchain sharding was not just theoretically possible but practically achievable at scale. Its EC-Schnorr multisignature approach for efficient PBFT was later referenced in multiple academic sharding papers. Scilla remains an important early example of formally verified smart contract language design.

---

Related Terms

• Sharding
• PBFT
• Smart Contract
• Harmony Whitepaper

---

Research

• Luu, L., et al. (2017). The Zilliqa Technical Whitepaper. Zilliqa Research.

— Primary whitepaper. Section 6 provides the security analysis; Section 7 reports testnet throughput.

• Castro, M., & Liskov, B. (1999). Practical Byzantine Fault Tolerance. USENIX OSDI 1999.

— Foundational PBFT paper; Zilliqa’s consensus is a multisignature-optimized variant of this system.

• Ankele, R., Khalid, A., Murphy, J., & Marnane, W. (2019). SoK: Sharding on Blockchain. Proceedings of the 1st ACM Conference on Advances in Financial Technologies.

— Survey that uses Zilliqa as a primary case study for production sharding design and implementation.