-
Centralisation on the Web has been a topic of increasing concern recently, with a
corresponding interest in decentralisation. In particular, there are concerns about
personal data in large volumes aggregated in the control of a single entity who is not the
subject of the data. Our vision is of a world where individuals are in control of their
own personal semantic data, on decentralised platforms following the FAIR TRADE
principles [
5
] { in particular, where the content of such data can be veri ed and trusted
by third parties wherever needed.
Centralisation of data can have global e ects. The potential in uence of Cambridge
Analytica on recent major political events was reported as a major scandal [
1
], for
example { made possible by the quantities of personal data held and controlled by a
single entity, Facebook. These concerns are not restricted solely to the digital domain
either. The loss of centralised records and distrust of non-centrally-held records led to
63 British citizens being deported, and up to 57 000 more being at risk of losing their
UK rights and residency [
2
], [
6
].
Issues such as these have prompted initiatives such as Solid [
4
], which aims to o er
personal Linked Data \pods", allowing users to hold and control their own data. From
another direction, the concept of blockchains [
7
] arose as a distributed ledger ; a shared
append-only record of immutable securely timestamped entries.
In [
5
], we presented \FAIR TRADE", an extension of the FAIR principles [
8
] for
data management to cover approaches to decentralised data. Findable, Accessible,
Interoperable, and Reusable are extended with TRusted, Autonomous, Distributed,
and dEcentralised. We present LinkChains as a FAIR TRADE platform for personal
Linked Data using blockchains to provide a trust backend. LinkChains o ers user
privacy and control, and granular data veri cation. A cryptographic hash of metadata
relating to personal data is stored in an immutable blockchain record, signed by a
trusted party. Such a hash serves as a pointer to the metadata location on decentralised
le storage platforms such as IPFS [
3
] capable of storing larger volumes of data more
easily and cheaply than most blockchain platforms. The immutability of the blockchain
record, and the nature of hashing, ensures the timestamp, immutability, and immunity
from faking of the IPFS data. An individual's Solid pod contains their actual data, as
Linked Data where possible, with pointers both to the blockchain and IPFS records.
Relevant hashes of data and metadata in each location can be collectively used to
verify the publication and untampered status of each.
LinkChains is thus a fully decentralised platform which supports individual hosting
and control of their personal data while maintaining the ability for external parties to
verify its contents where needed. Further extension to LinkChains involves the use of
Merkle trees of subsets of the data, as ne-grained as terms or quads. The tree data
ensures that queries can return veri able proofs of data at the given granularity {
we refer to this augmented query architecture as MerQL. The strength of the MerQL
approach is with federation; proofs can be included with individual data items included
in results aggregated from federated, sources.
[
5
] shows how LinkChains can support FAIR TRADE data management. In brief,
the novel principles (TRADE) are supported since LinkChains data can be Trusted
via the blockchain records and hashing, Autonomous due to individual control of
personal data, and Distributed and dEcentralised due to the underlying Solid, distributed
storage, and blockchain technology. LinkChains provides a research environment for
exploring Linked Data and blockchains in general, and how notions of trust, identity,
and decentralisation interact with each other, as well as enabling the implementation
and study of the implications of the vision we opened with: of a world where
individuals are in control of their own personal semantic data, on trusted FAIR TRADE
decentralised platforms.