LibraryData and agents2019Design paperCorpus record
Ocean Protocol: A Decentralized Substrate for AI Data and Services
Ocean. Trent McConaghy and the Ocean Protocol team.
Ocean's technical paper for publishing, pricing and consuming data services with on-chain access control and off-chain storage. The data does not sit inside the chain. The permission and the payment do.
Ocean's paper describes a marketplace where data stays where it is, and the chain records the access agreement, the payment, and the pointer, rather than the file itself.
- Evidence
- Primary paper
- Re-measured
- No
- Assumptions
- 4
- Records linked
- 0
Code
Source, repository, and what is actually hosted
- Ocean technical white paperhttps://oceanprotocol.com/tech-whitepaper.pdf
- The source URL is not a GitHub repository. No code remote is invented.
- Full text is not mirrored. Public availability is not a licence to copy the file onto this site.
- Add to a reading list
01 Claim ledger
What the paper is allowed to say
Each row is a sentence already in the study. The status is the same on every row: a model claim, not a live measurement.
Claim 01 · Paper model
Data does not belong on the ledger
Files are large, private, and often illegal to publish. The paper puts a data token and an access-control story on-chain, and leaves the bytes off-chain.
Claim 02 · Paper model
A data token is a right, not the dataset
Holding or spending the token is how a consumer is entitled to the service. The token is not a container. Anyone who describes it as 'the data' has misread the design.
Claim 03 · Paper model
Providers run the compute or the download
A publisher advertises a service. A consumer pays. A provider node, in the paper, is what actually hands over access. Availability is a service-level fact.
Claim 04 · Paper model
Compute-to-data is the privacy move
Later Ocean writing emphasises sending the algorithm to the data rather than the data to the buyer. The core paper's contribution is the split between access control and storage. Say which version you are citing.
Claim 05 · Paper model
Markets in data repeat ordinary market problems
Quality, exclusivity and resale are not solved by a token. The paper gives a way to charge for access. It does not give a way to unsee a file.
02 Three cuts
Observation, model, falsifier
A desk does not stop at the summary. Each claim is cut three ways, using only this study's own assumptions and checks. Nothing here is a new figure.
01 Data does not belong on the ledger
Observation
What the study says
Files are large, private, and often illegal to publish. The paper puts a data token and an access-control story on-chain, and leaves the bytes off-chain.
Model
What has to hold
At least one provider is online and willing to honour a valid token.
Falsifier
What would retire it
Where are the bytes, and who can switch them off?
02 A data token is a right, not the dataset
Observation
What the study says
Holding or spending the token is how a consumer is entitled to the service. The token is not a container. Anyone who describes it as 'the data' has misread the design.
Model
What has to hold
At least one provider is online and willing to honour a valid token.
Falsifier
What would retire it
What does the token entitle the holder to, in one sentence?
03 Providers run the compute or the download
Observation
What the study says
A publisher advertises a service. A consumer pays. A provider node, in the paper, is what actually hands over access. Availability is a service-level fact.
Model
What has to hold
At least one provider is online and willing to honour a valid token.
Falsifier
What would retire it
What is published on-chain about the buyer and the dataset?
04 Compute-to-data is the privacy move
Observation
What the study says
Later Ocean writing emphasises sending the algorithm to the data rather than the data to the buyer. The core paper's contribution is the split between access control and storage. Say which version you are citing.
Model
What has to hold
Buyers understand they are purchasing access under a policy, not a unique uncopyable object.
Falsifier
What would retire it
What is published on-chain about the buyer and the dataset?
05 Markets in data repeat ordinary market problems
Observation
What the study says
Quality, exclusivity and resale are not solved by a token. The paper gives a way to charge for access. It does not give a way to unsee a file.
Model
What has to hold
At least one provider is online and willing to honour a valid token.
Falsifier
What would retire it
What does the token entitle the holder to, in one sentence?
03 Sequence
One action, as an operating tape
- 01A publisher registers a dataset or a service and receives a data token that represents access.
- 02The bytes stay in storage the publisher or a provider controls.
- 03A consumer buys or is transferred the right the token represents.
- 04The consumer presents that right to a provider node, which checks it and serves the download or runs the job.
- 05Payment and the access event can be recorded. The file is not reconstructed from the chain.
04 Load-bearing
The argument, and where a pitch drops it
Off-chain delivery reintroduces the operator
The cut
The chain can prove a token was paid. It cannot by itself force a server to send the file. The paper's honesty depends on providers who can be replaced or whose reputation matters. A design review should spend its time on that server, not on the token symbol.
Why it carries weight
If this cut is skipped, the paper's name is being used without the mechanism that makes the name mean anything.
Where it is dropped
The paper is not a privacy proof. A provider that can read the file can copy the file, unless a later mechanism truly prevents that.
A copied file has escaped
The cut
Once bytes are delivered, the token cannot claw them back. Pricing power exists before delivery. The paper is a vending mechanism. Confidential compute, if used, is a later constraint on what the buyer receives. It is not automatic.
Why it carries weight
If this cut is skipped, the paper's name is being used without the mechanism that makes the name mean anything.
Where it is dropped
The paper is not a privacy proof. A provider that can read the file can copy the file, unless a later mechanism truly prevents that.
Metadata will leak more than the architecture admits
The cut
Titles, schemas and who bought access are often the sensitive part. An on-chain marketplace that publishes them has made a privacy decision. The paper should be checked for what it puts in public logs.
Why it carries weight
If this cut is skipped, the paper's name is being used without the mechanism that makes the name mean anything.
Where it is dropped
The paper is not a privacy proof. A provider that can read the file can copy the file, unless a later mechanism truly prevents that.
05 Register
What has to be true
Network · Not re-measured
At least one provider is online and willing to honour a valid token.
Model · Not re-measured
The pointer in the registry still resolves. A token whose URL rotted is a souvenir.
Model · Not re-measured
Buyers understand they are purchasing access under a policy, not a unique uncopyable object.
Data · Not re-measured
Whatever is logged on-chain is acceptable to the data's subjects.
06 Divergence
What happened after the paper
Ocean deployed data tokens and a marketplace, and later emphasised compute-to-data. Adoption was a marketplace problem as much as a protocol problem: listing a dataset does not create a buyer. The paper remains a clear statement that the chain should mediate rights and payments, not become the disk.
A later client, parameter set, or reward formula is a different object. Cite this paper for the mechanism. Cite a primary release for the network. This desk has not re-run the proof.
07 Pre-mortem
What to check before you use the idea
- 0 of 4 marked on this browser. A mark is a reading note, not a pass, a rating, or a recommendation.
08 Anatomy
The paper, in the order a builder needs
The problem it names
Useful datasets are too large and too sensitive to paste into a contract, and a marketplace that holds the files becomes the new platform. Ocean's design separates the asset pointer, the access contract and the place the bytes actually live.
What the design proposes
- A data token or access contract represents permission, not the file itself.
- Providers serve the bytes after the chain says the buyer is entitled.
- Curated registries and pricing mechanisms are how the paper imagines discovery.
How the mechanism is specified
- The chain can show who paid. It cannot, by itself, show that the file was worth buying. Quality is a market and a review problem.
- Compute-to-data, in later Ocean writing, is the idea that a buyer runs a job where the data sits rather than downloading it. Treat that as a specific feature with its own trust in the compute provider.
- Storage backends remain a dependency. Decentralising the permission does not decentralise a broken disk.
What this page does not treat as proven
- The paper is not a privacy proof. A provider that can read the file can copy the file, unless a later mechanism truly prevents that.
- We do not describe token price or marketplace volume.
- Later Ocean versions change the token mechanics. The architectural split is what this page keeps.
Why the desk still reads it
Agent ventures that 'sell data' need this split: the entitlement, the bytes, and the job that runs on the bytes. Collapsing all three into one server is a normal company. It should be described as one.
09 Lexicon
Terms, opened into the record
- Data token
- A token that represents a right to a data service. It does not contain the data.
- Provider
- The node that stores or computes and that honours a valid access right.
- Compute-to-data
- Sending the algorithm to the dataset so the buyer receives a result rather than a copy of the file.
- Registry
- The on-chain record that points at a service and its terms. The pointer can rot.
10 Repository
Every linked record on this page
Underlined words open a page that already exists: a concept, a protocol profile, a failure record, or another paper. If a word is not underlined, this desk does not have a record for it.
No repository match in this text.
This is Blockchain Lab's reading of a public design paper. It is not the paper, not a copy of it, and not an offer of tokens, equity, custody or a partnership. Later network behaviour can diverge from the text. Nothing here is investment, legal or technical advice.
Research status: Design paper. Last reviewed: 1 October 2026. This is a reading of a public paper, not investment, legal or security advice.