Re: [Hyperledger Project TSC] Cello conversations + Community collaboration thoughts

Baohua Yang

Dear Brian, Chris and Dan

Thanks for all your insightful suggestions!

Agree, dashboard is not the key point for Cello, as illustrated in the architecture

As an initial effort, the core of Cello is to target a backend BaaS engine. And it provides two kinds of front-end RESTful APIs for 1) chain-users and 2) system operators. 

To demo the capability, we integrate an operational dashboard for system operators, which is used to create/delete/operate multiple chains. This is of a different goal from what our Explorer does today (to "view/query" status of a chain as a "user-friendly web application"). At the same time, you may notice that there's no default chain-user dashboard in Cello yet, because we'd like to think it's up to the end-consumers, who may want to adopt their own dashboards in different scenarios.

Today, we're very glad to find that Explorer can be a good candidate as the chain-user dashboard, and Chaintool as the chaincode pkg tool for Cello, as practiced by one of Cello's consumers. This opens the possibility to make some "suite", which can be recommended to Hyperledger users after more verifications. To help that, I just got a patchset merged recently to containerize the Explorer as the first step.

It would be nice if Cello can attract more people into the community by taking Hyperledger easily, to adopt fabric, sawtoothlake, iroha and even more as the underly ledger, to use projects like Explorer and chaintool as development tools. 

IMHO, besides Cello as the BaaS effort, there are still lots of interesting topics to exploit, e.g., Blockchain+BigData, Blockchain+Cognitive Analysis. We hope to encourage more such ideas and discussions to make the community more energetic.

This is gonna a long-term evolution and need every Hyperledger contributor's help!

Thanks again for all the advice and have a great weekend!

On Sat, Dec 10, 2016 at 12:53 AM, Middleton, Dan via hyperledger-tsc <hyperledger-tsc@...> wrote:

I think we want a big tent while at the same time not proliferating narrow utilities tied to existing projects as separate projects unto themselves.


For example, I do not think it would make sense to spin out the ‘sawtooth cluster’ command as a separate project even though it is a useful way to spawn a network of validators.


I initially felt that Cello would stand a better chance of success with a broader scope and a broader team, hence merging with explorer.


The committed scope, though, to expand beyond launching Fabric dockers to include the other infrastructure projects (Iroha and Sawtooth Lake) seems like it may rise to the level of a distinct project.


I’m inclined towards more rather than fewer projects, but I would appreciate hearing views from the rest of the TSC with considerations on repo clutter, dilution, overhead, etc.





From: [mailto:hyperledger-tsc-bounces@...] On Behalf Of Brian Behlendorf via hyperledger-tsc
Sent: Friday, December 09, 2016 11:29
To: Christopher Ferris <chris.ferris@...>
Subject: Re: [Hyperledger Project TSC] Cello conversations + Community collaboration thoughts


I see Explorer and Cello as pretty different, actually, with nearly(?) no overlap in functionality other than both expose a web interface to an end user, so let's bring Cello in and let both teams figure out over time what moves where.


On December 9, 2016 11:13:30 AM EST, Christopher Ferris <chris.ferris@...> wrote:

+1 I agree, I would suggest we proceed and where it makes sense, we start leveraging more of explorer rather than reinvent the wheel, and where it makes sense we should see function migrate from Cello to Explorer etc.




On Fri, Dec 9, 2016 at 7:58 AM, Brian Behlendorf via hyperledger-tsc <> wrote:

On 12/08/2016 11:54 PM, Baohua Yang via hyperledger-tsc wrote:

Dear TSC members


After last TSC meeting, Dan, Pardha, Satheesh and I make some conversations on Cello's proposal and the collaborations in community, with more and more projects joining Hyperledger today. 


We bring some interesting thoughts together and wish share out for boarder discussions.


(Feel free to replenish if i missed some points.)


1. What kind of project are welcome to Hyperledger?

Hyperledger community should be open to any kind of technical contributions, especially those can fill new scopes, meet different requirements, and work well with existing projects.


Cello is exactly filling the gap of BaaS (Blockchain as a Service) requirement, and its proposed scope will let it work very well with the existing projects, e.g., fabric, sawtoothlake, iroha, blockchain-explorer, chaintool and sdk. It would be nice to consider Cello as the initial effort towards Hyperledger BaaS platform.


We welcome more such kind of collaboration potentials between different projects, under the same umbrella, while remaining the users' flexibility to choose what projects can meet their requirement best. 

My personal read of the conversation: I didn't sense any resistance to the idea of Cello, whether a Blockchain-as-a-service application would belong at Hyperledger, amongst the TSC or the people at the HackFest or anywhere else.  This seems to fill an important need when it comes to building and managing the "dev ops" side of running a chain.  I think the question with respect to Explorer was simply whether these belonged as two separate projects from day one, or whether the scope of Explorer could be expanded to also include Cello.   I think this broader question is still interesting - how big is our big tent? - but w/r/t to Cello I'd say don't worry about it.  Others?

My personal take: we should bring it in and not worry up front about forcing two different codebases (and potentially two different communities) to integrate before getting started.  If it makes sense technically or from a product perspective, it'll happen.  More granularity between projects early on means a faster way to run experiments (as Dan Middleton put it at the member summit) and keep an eye to building a more scalable community.

2. How to encourage better collaborations inside the Community?

To encourage more collaboration between various projects from different teams with different programming languages, we may bring up the concept of "suite".


A suite is a collection of Hyperledger projects, to achieve some open-to-use solution for specific scenarios, e.g., for Cloud scenario, for BigData scenario, for stock-trading scenario. This can help mitigate the technical challenge for users adoptions, and also help develop teams of the same suite to work more close with each other.


After enough practical adoptions, we may provide various community-recommended suites, just as the Linux, OpenStack and BigData communities do.

Right.  There could will be some projects under the Hyperledger banner which are aggregations of the output of other projects, plus maybe new code too.  It could be, for instance, that over time, Fabric makes sense to divide into a handful of smaller projects to make it easier to build a community around each, with well-defined APIs in between, and then the "Fabric" project would aggregate those pieces together along with an installer and other glue code and be the larger combined work.  Likewise, Cello might incorporate Explorer's code and functionality as one screen tab among several within Cello - no problems at all with that.




Would be glad if these ideas can attract more thoughts.


Look forward to hearing more insightful comments!





Best wishes!

Baohua Yang

hyperledger-tsc mailing list



Brian Behlendorf
Executive Director, Hyperledger
Twitter: @brianbehlendorf

hyperledger-tsc mailing list


Sent from my Android device with K-9 Mail. Please excuse my brevity.

hyperledger-tsc mailing list

Best wishes!

Baohua Yang

Join to automatically receive all group messages.