Topics

Hyperledger Cactus Academic Paper


Rafael André Pestana Belchior
 

Hello everyone.
I believe it would be extremely beneficial to Hyperledger Cactus to have a scientific paper detailing Cactus' mission, design, and implementation, written by maintainers and contributors. As a reference, check Fabric's paper: https://dl.acm.org/doi/10.1145/3190508.3190538 (over 13k reads, hundreds of citations).

In fact, after the scientific paper on Fabric was published, the usage of that blockchain by academics and the industry skyrocketed. Moreover, the timing is good: academic and industry research on interoperability problems is emerging at a very fast pace. This is a good opportunity to catch the next wave.

As confirmed by some maintainers, the academic paper is compatible with Cactus' mission. People interested in contributing please ping me so I can give you write access to the paper, written in the collaborative platform Overleaf.

Possible conferences to submit:
DSN ( www.dsn.org - dec)
ATC ( https://www.usenix.org/conference/atc21 - january)

Cheers,

--
Rafael Belchior
Ph.D. student in Computer Science and Engineering, Blockchain - Técnico
Lisboa
https://rafaelapb.github.io/
https://www.linkedin.com/in/rafaelpbelchior/


clive boulton
 

Hi Rafael, 

I'm also interested in contributing to the interoperability mission. More towards capabilities-based software security for trusted plugins and low latency connections. I signed up for Overleaf using clive.boulton@... 

Thank you,
LF @cliveb
    


On Thu, Aug 6, 2020 at 9:19 AM Rafael André Pestana Belchior <rafael.belchior@...> wrote:
Hello everyone.
I believe it would be extremely beneficial to Hyperledger Cactus to have
a scientific paper detailing Cactus' mission, design, and
implementation, written by maintainers and contributors. As a reference,
check Fabric's paper: https://dl.acm.org/doi/10.1145/3190508.3190538
(over 13k reads, hundreds of citations).

In fact, after the scientific paper on Fabric was published, the usage
of that blockchain by academics and the industry skyrocketed. Moreover,
the timing is good: academic and industry research on interoperability
problems is emerging at a very fast pace. This is a good opportunity to
catch the next wave.

As confirmed by some maintainers, the academic paper is compatible with
Cactus' mission. People interested in contributing please ping me so I
can give you write access to the paper, written in the collaborative
platform Overleaf.

Possible conferences to submit:
DSN ( www.dsn.org  - dec)
ATC ( https://www.usenix.org/conference/atc21 - january)

Cheers,

--
Rafael Belchior
Ph.D. student in Computer Science and Engineering, Blockchain - Técnico
Lisboa
https://rafaelapb.github.io/
https://www.linkedin.com/in/rafaelpbelchior/




shingo_fujimoto@fujitsu.com
 

Hello friends,

 

I agree with you to write an academic paper for getting feedback from non-developers.

Please include me as your collaborator :-)

 

My little concern is expected fields of contents are too wide in this thread.

I might suggest starting with the table of contents to align the thoughts into single main topic.

 

Best,

Shingo

 

---

Shingo Fujimoto

Senior Researcher, Blockchain Research Center, Security Laboratory,

FUJITSU LABORATORIES LTD

 

From: cactus@... <cactus@...> On Behalf Of clive boulton
Sent: Friday, August 7, 2020 4:58 AM
To: cactus@...
Subject: Re: [Hyperledger Cactus] Hyperledger Cactus Academic Paper

 

Hi Rafael, 

 

I'm also interested in contributing to the interoperability mission. More towards capabilities-based software security for trusted plugins and low latency connections. I signed up for Overleaf using clive.boulton@... 

 

Thank you,

LF @cliveb

    

 

On Thu, Aug 6, 2020 at 9:19 AM Rafael André Pestana Belchior <rafael.belchior@...> wrote:

Hello everyone.
I believe it would be extremely beneficial to Hyperledger Cactus to have
a scientific paper detailing Cactus' mission, design, and
implementation, written by maintainers and contributors. As a reference,
check Fabric's paper: https://dl.acm.org/doi/10.1145/3190508.3190538
(over 13k reads, hundreds of citations).

In fact, after the scientific paper on Fabric was published, the usage
of that blockchain by academics and the industry skyrocketed. Moreover,
the timing is good: academic and industry research on interoperability
problems is emerging at a very fast pace. This is a good opportunity to
catch the next wave.

As confirmed by some maintainers, the academic paper is compatible with
Cactus' mission. People interested in contributing please ping me so I
can give you write access to the paper, written in the collaborative
platform Overleaf.

Possible conferences to submit:
DSN ( www.dsn.org  - dec)
ATC ( https://www.usenix.org/conference/atc21 - january)

Cheers,

--
Rafael Belchior
Ph.D. student in Computer Science and Engineering, Blockchain - Técnico
Lisboa
https://rafaelapb.github.io/
https://www.linkedin.com/in/rafaelpbelchior/



hmontgomery@us.fujitsu.com
 

Hi Fujimoto-san,

Have you read the Fabric paper?  If you haven't, it's here:  https://arxiv.org/pdf/1801.10228.pdf.

This paper is written in the classic academic systems security style, which is what we should mimic if we want to maximize exposure in the academic community (and acceptance to a good conference).  I think coming up with an outline might be a good idea, but we probably don't need a table of contents.  The Fabric paper outline seems good, so we can base the outline on the Fabric paper, with an introduction, a background section, architecture section(s), and small applications, related work, and conclusion sections.  The only thing I think we might want to add is a "security" section, which details our "threat model" (really just the trust model) for the system and proves that if this is satisfied, transactions occur.

What do you think about this?  

Thanks,
Hart 


From: cactus@... <cactus@...> on behalf of shingo_fujimoto@... <shingo_fujimoto@...>
Sent: Thursday, August 6, 2020 5:03 PM
To: cactus@... <cactus@...>
Subject: Re: [Hyperledger Cactus] Hyperledger Cactus Academic Paper
 

Hello friends,

 

I agree with you to write an academic paper for getting feedback from non-developers.

Please include me as your collaborator :-)

 

My little concern is expected fields of contents are too wide in this thread.

I might suggest starting with the table of contents to align the thoughts into single main topic.

 

Best,

Shingo

 

---

Shingo Fujimoto

Senior Researcher, Blockchain Research Center, Security Laboratory,

FUJITSU LABORATORIES LTD

 

From: cactus@... <cactus@...> On Behalf Of clive boulton
Sent: Friday, August 7, 2020 4:58 AM
To: cactus@...
Subject: Re: [Hyperledger Cactus] Hyperledger Cactus Academic Paper

 

Hi Rafael, 

 

I'm also interested in contributing to the interoperability mission. More towards capabilities-based software security for trusted plugins and low latency connections. I signed up for Overleaf using clive.boulton@... 

 

Thank you,

LF @cliveb

    

 

On Thu, Aug 6, 2020 at 9:19 AM Rafael André Pestana Belchior <rafael.belchior@...> wrote:

Hello everyone.
I believe it would be extremely beneficial to Hyperledger Cactus to have
a scientific paper detailing Cactus' mission, design, and
implementation, written by maintainers and contributors. As a reference,
check Fabric's paper: https://dl.acm.org/doi/10.1145/3190508.3190538
(over 13k reads, hundreds of citations).

In fact, after the scientific paper on Fabric was published, the usage
of that blockchain by academics and the industry skyrocketed. Moreover,
the timing is good: academic and industry research on interoperability
problems is emerging at a very fast pace. This is a good opportunity to
catch the next wave.

As confirmed by some maintainers, the academic paper is compatible with
Cactus' mission. People interested in contributing please ping me so I
can give you write access to the paper, written in the collaborative
platform Overleaf.

Possible conferences to submit:
DSN ( www.dsn.org  - dec)
ATC ( https://www.usenix.org/conference/atc21 - january)

Cheers,

--
Rafael Belchior
Ph.D. student in Computer Science and Engineering, Blockchain - Técnico
Lisboa
https://rafaelapb.github.io/
https://www.linkedin.com/in/rafaelpbelchior/



@DorisBenda
 

Hello, academic paper enthusiasts,

Traditionally, academic papers are heavily rated and evaluated based on their novel contribution during the peer review process. Therefore,
among the sections introduction, background, architecture, related work ... (which can be loosely based on the white paper or talks given by cactus members), there should be a section called numerical evaluation, benchmarking or experimental section which uses graphs to derive a novel contribution or affirms a novel hypothesis. The "5.2 Experiments" section in the Fabric paper is a good example. They evaluate the performance of Fabric in terms of throughput based on several parameters that are varied. Since Fabric was new at that time, this performance evaluation was their novel contribution. I feel, that the "novel contribution" and the related sections will need some work and discussion among people because so far there is nothing in the white paper that can be recycled for it.

Just to clarify, this will be a conference paper in the end or can it also be journal paper?


Rafael André Pestana Belchior
 

Hello all,

First of all, thank you all for your interest! @Peter, @Clive, @Hart, @Fujimoto-san, great to have you onboard! I've added you to the shared document where we can start discussing our paper. 

@Hart,

I totally agree. The outline for the paper (as well as the "story" we want to transmit) are crucial points. I think we should meet and discuss an initial outline, detailing the strong points of Cactus. I've added a preliminary outline, present in the shared Overleaf project.

Given this, shall we discuss a preliminary outline (maybe counting with the whitepaper as a reference) on the Cactus Contributors Meeting?

@Doris,

Agreed. I think the paper needs to show what Cactus brings to the academic and industry communities, and it should of course contain a solid evaluation. Some good questions that the evaluation responds to are: what is the degree of decentralization of a Cactus Consortium? What is the performance penalty for mediating transactions across 2 blockchains with Cactus? And with 3? How does the geographical distance between Cactus' nodes affect the consortium performance, etc. We need to discuss this as a team, along with @Peter.

I think the novel contribution that Fabric brought was not its performance evaluation methodology, but rather being the first (if i'm not mistaken) distributed operating system for permissioned blockchains, tailored for enterprises. Their strongest innovation, in my POV, is the execute-order-validate paradigm, which allowed for much better tx throughput than public blockchains.

Conversely, Cactus has a lot of strong points that differentiates it from current academic work - it can pave its way to becoming the status quo for enterprise interoperability solutions.

Conference vs Journal: in my view, I would like to submit on the place that brings the most visibility to this project. In light of this, we should submit to a top conference or a top journal. My research group advised me to write my papers in a way I express all the ideas and concepts, without summarizing anything. Then, depending on the submission site, the article may be shortened. This allows us to be flexible and keep doors opened regarding the submission. 

Looking to hearing from you.

Cheers,

Rafael

A 2020-08-07 11:17, doris.benda via lists.hyperledger.org escreveu:

Hello, academic paper enthusiasts,

Traditionally, academic papers are heavily rated and evaluated based on their novel contribution during the peer review process. Therefore,
among the sections introduction, background, architecture, related work ... (which can be loosely based on the white paper or talks given by cactus members), there should be a section called numerical evaluation, benchmarking or experimental section which uses graphs to derive a novel contribution or affirms a novel hypothesis. The "5.2 Experiments" section in the Fabric paper is a good example. They evaluate the performance of Fabric in terms of throughput based on several parameters that are varied. Since Fabric was new at that time, this performance evaluation was their novel contribution. I feel, that the "novel contribution" and the related sections will need some work and discussion among people because so far there is nothing in the white paper that can be recycled for it.

Just to clarify, this will be a conference paper in the end or can it also be journal paper?


--
Rafael Belchior
Ph.D. student in Computer Science and Engineering, Blockchain - Técnico Lisboa
https://rafaelapb.github.io/
https://www.linkedin.com/in/rafaelpbelchior/


shingo_fujimoto@fujitsu.com
 

Hi all,

 

I believe this is a nice workout to clarify our design issue remained.

 

When we finished v1 implementation, I am okay with writing a full paper such in near future, but I suggest you start writing some short conference papers, which discuss the specific topic at the working progress stage.

 

I tried to figure our following potential topics:

-            Threats on blockchain integrated service

-            Ledger migration for long-term persistent

-            Introducing secure environment for blockchain integrated service

 

 

We might discuss at next call for potential topics in such papers.

 

It’s summer and a good time for writing something  :-)

 

BR,

Shingo

 

From: cactus@... <cactus@...> On Behalf Of Rafael Andre Pestana Belchior
Sent: Saturday, August 8, 2020 3:44 AM
To: cactus@...
Subject: Re: [Hyperledger Cactus] Hyperledger Cactus Academic Paper

 

Hello all,

First of all, thank you all for your interest! @Peter, @Clive, @Hart, @Fujimoto-san, great to have you onboard! I've added you to the shared document where we can start discussing our paper. 

@Hart,

I totally agree. The outline for the paper (as well as the "story" we want to transmit) are crucial points. I think we should meet and discuss an initial outline, detailing the strong points of Cactus. I've added a preliminary outline, present in the shared Overleaf project.

Given this, shall we discuss a preliminary outline (maybe counting with the whitepaper as a reference) on the Cactus Contributors Meeting?

@Doris,

Agreed. I think the paper needs to show what Cactus brings to the academic and industry communities, and it should of course contain a solid evaluation. Some good questions that the evaluation responds to are: what is the degree of decentralization of a Cactus Consortium? What is the performance penalty for mediating transactions across 2 blockchains with Cactus? And with 3? How does the geographical distance between Cactus' nodes affect the consortium performance, etc. We need to discuss this as a team, along with @Peter.

I think the novel contribution that Fabric brought was not its performance evaluation methodology, but rather being the first (if i'm not mistaken) distributed operating system for permissioned blockchains, tailored for enterprises. Their strongest innovation, in my POV, is the execute-order-validate paradigm, which allowed for much better tx throughput than public blockchains.

Conversely, Cactus has a lot of strong points that differentiates it from current academic work - it can pave its way to becoming the status quo for enterprise interoperability solutions.

Conference vs Journal: in my view, I would like to submit on the place that brings the most visibility to this project. In light of this, we should submit to a top conference or a top journal. My research group advised me to write my papers in a way I express all the ideas and concepts, without summarizing anything. Then, depending on the submission site, the article may be shortened. This allows us to be flexible and keep doors opened regarding the submission. 

Looking to hearing from you.

Cheers,

Rafael

A 2020-08-07 11:17, doris.benda via lists.hyperledger.org escreveu:

Hello, academic paper enthusiasts,

Traditionally, academic papers are heavily rated and evaluated based on their novel contribution during the peer review process. Therefore,
among the sections introduction, background, architecture, related work ... (which can be loosely based on the white paper or talks given by cactus members), there should be a section called numerical evaluation, benchmarking or experimental section which uses graphs to derive a novel contribution or affirms a novel hypothesis. The "5.2 Experiments" section in the Fabric paper is a good example. They evaluate the performance of Fabric in terms of throughput based on several parameters that are varied. Since Fabric was new at that time, this performance evaluation was their novel contribution. I feel, that the "novel contribution" and the related sections will need some work and discussion among people because so far there is nothing in the white paper that can be recycled for it.

Just to clarify, this will be a conference paper in the end or can it also be journal paper?

 

--

Rafael Belchior
Ph.D. student in Computer Science and Engineering, Blockchain - Técnico Lisboa
https://rafaelapb.github.io/
https://www.linkedin.com/in/rafaelpbelchior/


clive boulton
 

Hi all, thinking about the academic paper and how Cactus differentiates from the status quo for enterprise interoperability solutions. Daniel Abadi has written about what is consistency anyway and discussed the difference between partitioned vs global consensus. Is it off-topic to consider Cactus strong points of coordination to avoid dealing with partition issues by engineering them away, such as by time-series engineering that (Google's) Spanner has implemented?  E.g. imagine 3 Fabric systems in series with 2 Cactus systems providing blockchain integration, is there a need for a global clock?

BR, 
Clive  


@DorisBenda
 

Hello Rafael, 

I am interested in the endeavor regarding the academic paper too. My Overleaf account address is: doris.benda@.... Thank you.


Best wishes,
Doris 


Rafael André Pestana Belchior
 

Hey Clive,

Do you mean if Cactus can provide linearizability on transactions spanning across several blockchains? 

If that's the question, I don't think so, as Cactus currently is a smart client that connects blockchains, but does not manage them (i.e., it is not a blockchain on its own, nor does it have a more complex transaction management tool). 

Cheers,

Rafael

A 2020-08-10 22:27, clive boulton escreveu:

Hi all, thinking about the academic paper and how Cactus differentiates from the status quo for enterprise interoperability solutions. Daniel Abadi has written about what is consistency anyway and discussed the difference between partitioned vs global consensus. Is it off-topic to consider Cactus strong points of coordination to avoid dealing with partition issues by engineering them away, such as by time-series engineering that (Google's) Spanner has implemented?  E.g. imagine 3 Fabric systems in series with 2 Cactus systems providing blockchain integration, is there a need for a global clock?
 
BR, 
Clive  


--
Rafael Belchior
Ph.D. student in Computer Science and Engineering, Blockchain - Técnico Lisboa
https://rafaelapb.github.io/
https://www.linkedin.com/in/rafaelpbelchior/


Rafael André Pestana Belchior
 

Added! :)

A 2020-08-11 02:47, doris.benda via lists.hyperledger.org escreveu:

Hello Rafael, 

I am interested in the endeavor regarding the academic paper too. My Overleaf account address is: doris.benda@.... Thank you.


Best wishes,
Doris 


--
Rafael Belchior
Ph.D. student in Computer Science and Engineering, Blockchain - Técnico Lisboa
https://rafaelapb.github.io/
https://www.linkedin.com/in/rafaelpbelchior/


clive boulton
 

Hi Rafael,

Spanning several blockchains, yes. Or even two blockchains spanning undersea cables (say Tokyo, New York), where latency can go haywire. I meant by dealing with partitions by avoiding them with a global clock, linearizability on transactions is attained. See Daniel Abadi's work on PACELC for NewSQL. Apache Beam also deals with known unknowns using GC and time-series to simplify stream processing.

Btw. Shingo had made a presentation to deal with trustworthy validators, also by implementing a Cactus system managed cache, I think. We very lightly discussed during Monday's Cactus contributors meeting. 

A coda. I was embedded at Intel working on this area: non-linearizability on transactions due to partitions spanning geographies. Variable latency could play just enough havoc to require engineering. 

On Tue, Aug 11, 2020 at 1:55 AM Rafael André Pestana Belchior <rafael.belchior@...> wrote:

Hey Clive,

Do you mean if Cactus can provide linearizability on transactions spanning across several blockchains? 

If that's the question, I don't think so, as Cactus currently is a smart client that connects blockchains, but does not manage them (i.e., it is not a blockchain on its own, nor does it have a more complex transaction management tool). 

Cheers,

Rafael

A 2020-08-10 22:27, clive boulton escreveu:

Hi all, thinking about the academic paper and how Cactus differentiates from the status quo for enterprise interoperability solutions. Daniel Abadi has written about what is consistency anyway and discussed the difference between partitioned vs global consensus. Is it off-topic to consider Cactus strong points of coordination to avoid dealing with partition issues by engineering them away, such as by time-series engineering that (Google's) Spanner has implemented?  E.g. imagine 3 Fabric systems in series with 2 Cactus systems providing blockchain integration, is there a need for a global clock?
 
BR, 
Clive  


--
Rafael Belchior
Ph.D. student in Computer Science and Engineering, Blockchain - Técnico Lisboa
https://rafaelapb.github.io/
https://www.linkedin.com/in/rafaelpbelchior/


Dilum Bandara
 

Hi All,

I would love to contribute to the research paper.
I'm currently working on blockchain migration and multi-blockchain data transfer. Also, have an interest in blockchain workload and smart contract security. 

@Rafael My Overleaf account is under dilum.bandara@...

I agree with most of the comments discussed here. Perhaps they can be addressed by focusing on 2 papers:
1. Initially a paper for a Demo Track of a reputed conference (A or A*) - Usually paper is 2-4 pg. They mostly look for implementations of interesting ideas. As far as we can showcase a good working prototype should be fine. As per demo, a pre-recorded video is now the new norm with COVID
2. Later, a full paper for a top conference (A or A*) - Usually 8-12 pg. Novel contribution (Cacus itself has several new aspects) plus a solid performance analysis is what we need.

Best regards,
Dilum


Rafael André Pestana Belchior
 

Hello Dilum,

Welcome onboard.

Cheers,

Rafael

A 2020-08-25 01:04, Dilum Bandara escreveu:

Hi All,

I would love to contribute to the research paper.
I'm currently working on blockchain migration and multi-blockchain data transfer. Also, have an interest in blockchain workload and smart contract security. 

@Rafael My Overleaf account is under dilum.bandara@...

I agree with most of the comments discussed here. Perhaps they can be addressed by focusing on 2 papers:
1. Initially a paper for a Demo Track of a reputed conference (A or A*) - Usually paper is 2-4 pg. They mostly look for implementations of interesting ideas. As far as we can showcase a good working prototype should be fine. As per demo, a pre-recorded video is now the new norm with COVID
2. Later, a full paper for a top conference (A or A*) - Usually 8-12 pg. Novel contribution (Cacus itself has several new aspects) plus a solid performance analysis is what we need.

Best regards,
Dilum


--
Rafael Belchior
Ph.D. student in Computer Science and Engineering, Blockchain - Técnico Lisboa
https://rafaelapb.github.io/
https://www.linkedin.com/in/rafaelpbelchior/


Rafael André Pestana Belchior
 

Hello All,

I have update the preliminary structure of the paper. Feedback is welcome! I created a group at Hyperledger Chat Cactus (with all who expressed their intentions to collaborate) to facilitate discussion about the paper. Please feel free to ping me if I haven't given you access.

Cheers,

Rafael

A 2020-08-25 08:25, Rafael André Pestana Belchior escreveu:

Hello Dilum,

Welcome onboard.

Cheers,

Rafael

A 2020-08-25 01:04, Dilum Bandara escreveu:

Hi All,

I would love to contribute to the research paper.
I'm currently working on blockchain migration and multi-blockchain data transfer. Also, have an interest in blockchain workload and smart contract security. 

@Rafael My Overleaf account is under dilum.bandara@...

I agree with most of the comments discussed here. Perhaps they can be addressed by focusing on 2 papers:
1. Initially a paper for a Demo Track of a reputed conference (A or A*) - Usually paper is 2-4 pg. They mostly look for implementations of interesting ideas. As far as we can showcase a good working prototype should be fine. As per demo, a pre-recorded video is now the new norm with COVID
2. Later, a full paper for a top conference (A or A*) - Usually 8-12 pg. Novel contribution (Cacus itself has several new aspects) plus a solid performance analysis is what we need.

Best regards,
Dilum


--
Rafael Belchior
Ph.D. student in Computer Science and Engineering, Blockchain - Técnico Lisboa
https://rafaelapb.github.io/
https://www.linkedin.com/in/rafaelpbelchior/


--
Rafael Belchior
Ph.D. student in Computer Science and Engineering, Blockchain - Técnico Lisboa
https://rafaelapb.github.io/
https://www.linkedin.com/in/rafaelpbelchior/


Benedikt Putz
 

Hey everyone,

Rafael messaged me about the plans for an academic paper for Cactus. I'd love to contribute to the paper regarding the security aspects of Cactus.

Quick intro: I'm doing research on the security aspects of (cross-chain) blockchain tech, most recently I've been working on security monitoring for Hyperledger Fabric. I'll be joining the call later today as well in case you have any questions.

Best regards,
Benedikt
-----------

Benedikt Putz

Ph.D. Student Blockchain & Cybersecurity

 

Chair of Information Systems

University of Regensburg


E-Mail: benedikt.putz@...