A Block Chain Solution for the Sri Lankan Land Registry — Part 1

Through this 2-part article, I’m explaining how my Final Year Research in Computer Science was commercialized (or productized). The research project titled “Permissioned Distributed Ledgers for Land Transactions” was presented and included in the proceedings of the Blockchain Forum at the 17th International Conference on Business Process Management held in Vienna in September 2019. The full paper could be obtained at https://link.springer.com/chapter/10.1007/978-3-030-30429-4_10. In this 1st Part, I’m explaining only those details of the research which are important towards understanding the final product; LRBSL. LRBSL which stands for “Land Registration Blockchain for Sri Lanka” has been explained in Part 2.

It is a known fact that the current manual land registration systems being practiced in Sri Lanka are not sufficiently effective and efficient. There are many negative implications of the present land registration systems, such as existence of a large number of unsolved land disputes, unclear tenure leading to land encroachment, fraudulent land transactions, missing entries or tampered registers, presence of multiple land laws in the country etc. The emergence of the concept of blockchain based land registries could be seen as a successful replacement not only for badly kept, mismanaged and/or corrupt land registries (e.g.: Honduras, Ghana) but also for countries with already well-functioning land registries (e.g.: Sweden, Georgia, Estonia).

As claimed by concepts of Distributed Systems, blockchain is one type of distributed ledger. Thus, the aim of the research was to provide a distributed ledger solution for the Sri Lankan land transaction scenario. Out of the two main types of distributed ledgers; unpermissioned (permissionless) and permissioned, it was inferred that a permissioned DLT solution would suit the Sri Lankan land transaction scenario, considering the similarity with the real scenario. A permissioned DLT network contains an authorized consortium of participants where distributed consensus is obtained through validation by a selected subset of ‘trusted validating nodes’. Through analyzing content of current folio system, three main types of validators per a land transaction, recognized by the Sri Lankan government (i.e. Registrar General‘s Department) could be identified. They are the (1) Regional Land Registrar on behalf of Regional Land Registry, (2) Notary and (3) Surveyor.

The high level architecture of the proposed Sri Lankan distributed land ledger is shown in Fig. 1.

Fig. 1 High level architecture of the proposed Sri Lankan distributed land ledger

• Seller will submit a transaction proposal indicating the sale of her land to buyer.
• Next, the submitted transaction will be sent to all validating nodes.
• Consensus messaging takes place among validating nodes.
• If the transaction is identified as valid, it will be added to the ledger of each validating node.
• Finally, buyer could receive the status of transaction from any of the validators.

It could be observed that some countries have implemented their blockchain based land registries to suit the title registration process while some countries have implemented their blockchain solution following the deed registration system. Since, majority of Divisional Secretariat divisions in Sri Lanka follow the deed registration system, the proposed solution preserves properties of the deed system.

Design of Optimal Land Ledger Content

The final solution presents optimal content for the ledger. Optimal ledger content was extracted from the current folio, by removing redundant details and adjusting attributes to suit a distributed ledger solution. The optimal ledger content included in the provided permissioned distributed land ledger consists of, Land ID, Location of land, Boundaries of land (N, E, W, S), Extent, Hash of plan, Hash of deed, Registration stamp duty, Owner, Remarks regarding transaction & Parent Land ID. For complete details on how the optimal land ledger content was derived, it is recommended to read the full paper.

Only the hash values of plan and deed were added to optimal ledger content, because in blockchain based land registries of all countries reviewed, only the hash value of data is embedded in the blockchain, while actual data which is generally large in size and confidential, is kept off-chain (in a traditional server). Through this move, content of a land transaction remains irrefutable.

Design of Transactions Against Distributed Land Ledger

Before explaining the proposed transactions, it is important to understand the internals of ledger subsystem.

The ledger (L) of a permissioned DLT platform (with blockchain as the underlying database structure) comprises of,

1. World state (W): Stores the state of the ledger at a given point in time.
2. Transaction log: Stores all transactions which have contributed towards current world state in blockchain B.

Fig. 2 Ledger sub system (Image source: https://hyperledger-fabric.readthedocs.io/en/release-2.0/ledger/ledger.html#the-ledger)

Fig. 3 demonstrates the taxonomy of transactions in the proposed land ledger solution where land transactions of the current system have been preserved. Here, “Land transactions”, refer only to change of ownership right of a particular piece of land between two parties.

Fig. 3 Taxanomy of Transactions in the proposed Sri Lankan distributed land ledger

queryLand query would facilitate clients to request details regarding a piece of land when the LandID is provided. Through this query, the world state W of a land in the land ledger is queried and latest values corresponding to the attributes in optimal ledger content for the land would be returned. queryAllLands query facilitates retrieving details of all lands of the ledger. Clients would be able to request for a change of ownership of an existing piece of land through changeLandOwner transaction.

Furthermore, a client would be able to request to split an existing land and register newly created lands with new owners, updated extents and boundaries through forkLand transaction. Two non-directly invocable functions; createLand and deleteLand have been introduced, in order to implement forkLand function. forkLand transaction checks whether the sum of the extents of the new lands is consistent with the extent of original land and whether boundaries of newly created lands do not overlap with each other. If so, it will invoke deleteLand transaction which will delete the original land from ledger, and createLand transaction will be invoked n times to create n number of new lands as requested by the transaction proposal. It is ensured that the solution preserves the consistency of the land ledger, before and after executing a set of transactions.

Although, this research has got away with the folio system, since all transaction details which have contributed towards current world state are available in the blockchain B, it is possible to obtain the pedigree/folio tree which corresponds to a particular land at any given time through getHistoryForLand query.

Up to what has been explained now, the proposed solution has not deviated from the current manual system in SL. The significant deviation of the proposed solution from the current system is at the architecture of the final solution model. There, all validators would have access to an island wide unified land ledger (Fig. 4). The final solution model was proposed after evaluating it against the current regional distributed land ledger (where each RLR holds an independent land ledger of its own lands), under different land transaction density conditions (workload distribution) and node failure conditions. The island wide solution records higher throughput, lower latency and tolerance for fail-stop conditions than the regional distributed land ledger. For more details on how the final architecture of the solution model was derived, please read the full paper.

Fig. 4 Island wide unified land ledger

To summarize; we propose a permissioned distributed land ledger solution for Sri Lanka where, the final solution presents optimal content for the ledger (extracted from the current folio), preserves current land transactions, has reassigned duties to state validators and has got away with the folio system while ensuring derivation of the pedigree/folio tree for a land at a given time. In addition, the extent of details accessible by validators in the present traditional land transaction scenario has been preserved in the proposed solution. The architecture of the final solution model is an island wide unified land ledger for which all validators have access to.

On an important note, Hyperledger Fabric was used to implement and evaluate the proposed permissioned distributed ledger solution throughout the research.

Please read Part-2 of the article which explains implementation details of LRBSL at,

A Block Chain Solution for the Sri Lankan Land Registry — Part 2