v5.0.0

What's new?

• #313 Computing logs are pushed to iexec-core API
• #291 Results in TEE mode are pushed using TEE-post-compute component - https://github.com/iExecBlockchainComputing/tee-worker-post-compute. Results can be pushed to IPFS or Dropbox.
• Results are pushed to IPFS with Standard & Tee mode to a dedicated result-proxy - https://github.com/iExecBlockchainComputing/iexec-result-proxy
• #315 Task result link is standardized and generic. It supports Ethereum, IPFS & Dropbox "storage" providers.
• #308 #307 #306 #294 Dapps should produce a computed.json file. Merged determinsm.iexec & callback.iexec into it.
• Requester params are now explicitly set (storage, encryption, ..) and properly handled by workers
• Full compatibility with new iExec Secret Managment Service - https://github.com/iExecBlockchainComputing/iexec-sms
• Compatibility with latest PoCo 5.1.0 smart contracts

v4.0.1

What's fixed?

• More resistance to unsync Ethereum nodes #289

v4.0.0

What's new?

• Native-token sidechain compatibility #261
• GPU support #269
• Optional log aggregation on remote workers for software debugging #254
• Task computing logs for iExec buidlers debugging #255

What's fixed?

• JWT issues #266

v3.2.0

What is new?

• Whitelisting: Workers can now be whitelisted on the core side, if the worker is not in the whitelist it will not be able to join the pool.
• Https: Workers can now connect to the core using https.

What is patched?

• The project has been updated to java 11 and web3j 4.3.0
• Internal refactoring to handle replicates update better
• Workers that ask for replicates too often will not get replicate
• Update workers configuration when they disconnect/reconnect
• If a worker's wallet is (almost) empty, it stops asking for replicates to the core.

Misc:

• The environment variable DATASET_FILENAME has been renamed to IEXEC_DATASET_FILENAME. Applications that use it should be updated.

v3.1.0

What's new?

• Full end-to-end encryption inside a Trusted Execution Environment (TEE) powered by Intel(R) SGX #212
• Implemented the Proof-of-Contribution (PoCo) Sarmenta's formula for a greater task dispatching

What's patched?

• A single FAILED replicate status when a completion is impossible
• WORKER_LOST is not set for replicates which are already FAILED
• Checks before many steps the contribution deadline is not reached
• Better handling of access token expiration #209
• Don't notify when contribute or reveal seems to fail since the tx could have been mined #205
• Restart when ethereum node is not available at start-up

v3.0.1

What's new?

• Retry result upload multiple times with delay if remote result repository is not responding
• Wait blockchain task status changes if node is not sync
• Dont update replicate if node is not sync
• Dont ask for new task if node is not sync

v3.0.0

This release contains the worker set of changes in the 3.0.0-alpha-X releases and some other features/fixes.
In summary this version introduces:

• A new architecture: the worker has been completely re-architectured from the version 2
• Latest PoCo use
• Better management of transaction with the ethereum blockchain
• Failover mechanisms: in case some workers are lost or restarted when working on a specific task, internal mechanisms will redistribute the task or use as much as possible the work performed by the lost / restarted workers
• iExec End-To-End Encryption with Secret Management Service (SMS): from this version, inputs and outputs of the job can be fully encrypted using the Secret Management Service.
• Decentralized oracle: If the result is needed by a smart contract, it is available directly on the blockchain.
• IPFS: data can be retrieved from IPFS and public results can be published on IPFS

For further informations on this version, please read our documentation

v3.0.0-alpha3

What's new?

• Enable overriding the node the worker will connect to (#127)
• Return chainReceipt when contribute/reveal fails for onchain verification (#147)
• The worker now can downloaded data from IPFS.
• Better recovery mechanism when the worker is shut down in the middle of a computation.
• Better management when the transaction receipt from the blockchain is null.
• Additional configuration if a worker is behind a proxy.
• General refactoring.
• Compatible with [email protected]

v3.0.0-alpha2

What's new?

• The worker will send to the scheduler the tx hash & the block number of transactions to keep track of what happened
• The worker will reboot after a scheduler reboot
• Improved STOMP websocket managment
• Compatible with [email protected]

v3.0.0-alpha1

This is the first alpha release of version 3.0.0