5 ms·

>> "Assuming the appliance remains secure" Makes me wonder what mathematics would prove this, or if it's assumed to be out of scope, and if so, what else is ou

by nxzero 10y ago

>> "Assuming the appliance remains secure" Makes me wonder what mathematics would prove this, or if it's assumed to be out of scope, and if so, what else is out of scope too.

lallysingh 10y ago
The article mentioned that the project includes formal verification. From a moment scanning Galois' web page: https://galois.com/research-development/software-correctness/ https://galois.com/research-development/software-correctness...
stelfer 10y ago
I'm pretty sure the Guardtime product is a Merkle tree of timestamps. I'm also pretty sure whatever protocol they use to maintain it would _not_ be immediately recognized as a "blockchain" protocol. Of course the protocol can be modeled and checked. The question is why does anyone think that this activity can apply to "blockchains" in general. But, welcome to the world of industrial research funding.
- ristoalas 10y ago
  (I work for Guardtime; I’m happy to answer any questions.) Some online resources might not make it very clear, but the KSI blockchain is indeed a blockchain; for example, new blocks are released periodically (1 block per second on average), and the new blocks contain hashes of the previous blocks, among other things. Then, blocks get distributed around so that users can verify their data against these blocks, etc.
  - stelfer 10y ago
    Most people associate "blockchains" with a protocol by which the "blocks" are inserted. Usually these things revolve around proof-of-work, proof-of-stake, proof-of-space, fairness criteria, etc. I think everyone gets that the Guardtime product is hashtree. From a formal verification perspective, the critical behavior is probably in the protocol, and most likely not the implementation of the datastructure or database.
    - ristoalas 10y ago
      One way to look at it is that KSI is a proof of stake blockchain where the stakeholders are (mostly) fixed. Often people would call that a private blockchain (or a consortium blockchain), to use the terminology from the following article: https://blog.ethereum.org/2015/08/07/on-public-and-private-blockchains/ https://blog.ethereum.org/2015/08/07/on-public-and-private-b...
      - unboxed_type 10y ago
        Thanks for that clarification! It is a great pleasure to talk with an knowledgeable person like you.
  - Jasper_ 10y ago
    > 1 block per second on average What stops a double-spend or other attacks like that in a 1-second timeframe? The 10 minute threshold of Bitcoin was chosen under belief that the network would be settled after roughly 10 minutes.
    - ristoalas 10y ago
      The speed comes from the fact that private and consortium blockchains tend not use Proof of Work (at least not as the only algorithm; you could mix and match, if you wanted to). Instead, such blockchains tend to use voting-based protocols (e.g. Proof of Stake), where a certain subset of nodes are trusted with voting rights, who together vote to approve the next block. In some systems you also have automatic punishments built in to incentivise correct behaviour (i.e., if someone attempts to fork a blockchain --- for example by voting for 2 conflicting forks -- and if other nodes see any evidence of that, the node who attempted the fork could automatically lose their rights as a trusted validator, and may also automatically lose a lot of money or face legal fees.)
javast98 10y ago
I'm from Guardtime - this is our secure appliance that we built with a team of engineers coming out of the US Military Ant-Tamper community: http://bit.ly/2cgXiDu http://bit.ly/2cgXiDu
- nxzero 10y ago
  Redirected link from bitly link above: https://media.voog.com/0000/0036/0235/files/Black_Lantern_data_sheet_201509.pdf https://media.voog.com/0000/0036/0235/files/Black_Lantern_da...
  - hossbeast 10y ago
    What's wrong with posting a bitly link?
    - alfiedotwtf 10y ago
      Transparency
digikata 10y ago
I was just reading the Wikipedia page on Linked timestamping yesterday. Though I don't know that this is what Galois & Guardtime are doing. https://en.wikipedia.org/wiki/Linked_timestamping https://en.wikipedia.org/wiki/Linked_timestamping

by nxzero 10y ago

>> "Assuming the appliance remains secure" Makes me wonder what mathematics would prove this, or if it's assumed to be out of scope, and if so, what else is out of scope too.

lallysingh 10y ago

The article mentioned that the project includes formal verification. From a moment scanning Galois' web page: https://galois.com/research-development/software-correctness/ https://galois.com/research-development/software-correctness...

stelfer 10y ago

I'm pretty sure the Guardtime product is a Merkle tree of timestamps. I'm also pretty sure whatever protocol they use to maintain it would _not_ be immediately recognized as a "blockchain" protocol. Of course the protocol can be modeled and checked. The question is why does anyone think that this activity can apply to "blockchains" in general. But, welcome to the world of industrial research funding.

ristoalas 10y ago
(I work for Guardtime; I’m happy to answer any questions.) Some online resources might not make it very clear, but the KSI blockchain is indeed a blockchain; for example, new blocks are released periodically (1 block per second on average), and the new blocks contain hashes of the previous blocks, among other things. Then, blocks get distributed around so that users can verify their data against these blocks, etc.
- stelfer 10y ago
  Most people associate "blockchains" with a protocol by which the "blocks" are inserted. Usually these things revolve around proof-of-work, proof-of-stake, proof-of-space, fairness criteria, etc. I think everyone gets that the Guardtime product is hashtree. From a formal verification perspective, the critical behavior is probably in the protocol, and most likely not the implementation of the datastructure or database.
  - ristoalas 10y ago
    One way to look at it is that KSI is a proof of stake blockchain where the stakeholders are (mostly) fixed. Often people would call that a private blockchain (or a consortium blockchain), to use the terminology from the following article: https://blog.ethereum.org/2015/08/07/on-public-and-private-blockchains/ https://blog.ethereum.org/2015/08/07/on-public-and-private-b...
    - unboxed_type 10y ago
      Thanks for that clarification! It is a great pleasure to talk with an knowledgeable person like you.
- Jasper_ 10y ago
  > 1 block per second on average What stops a double-spend or other attacks like that in a 1-second timeframe? The 10 minute threshold of Bitcoin was chosen under belief that the network would be settled after roughly 10 minutes.
  - ristoalas 10y ago
    The speed comes from the fact that private and consortium blockchains tend not use Proof of Work (at least not as the only algorithm; you could mix and match, if you wanted to). Instead, such blockchains tend to use voting-based protocols (e.g. Proof of Stake), where a certain subset of nodes are trusted with voting rights, who together vote to approve the next block. In some systems you also have automatic punishments built in to incentivise correct behaviour (i.e., if someone attempts to fork a blockchain --- for example by voting for 2 conflicting forks -- and if other nodes see any evidence of that, the node who attempted the fork could automatically lose their rights as a trusted validator, and may also automatically lose a lot of money or face legal fees.)

javast98 10y ago

I'm from Guardtime - this is our secure appliance that we built with a team of engineers coming out of the US Military Ant-Tamper community: http://bit.ly/2cgXiDu http://bit.ly/2cgXiDu

nxzero 10y ago
Redirected link from bitly link above: https://media.voog.com/0000/0036/0235/files/Black_Lantern_data_sheet_201509.pdf https://media.voog.com/0000/0036/0235/files/Black_Lantern_da...
- hossbeast 10y ago
  What's wrong with posting a bitly link?
  - alfiedotwtf 10y ago
    Transparency

digikata 10y ago

I was just reading the Wikipedia page on Linked timestamping yesterday. Though I don't know that this is what Galois & Guardtime are doing. https://en.wikipedia.org/wiki/Linked_timestamping https://en.wikipedia.org/wiki/Linked_timestamping