Hyperledger Foundation Hyperledger Global Forum 2020, 9 Mar 2020

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Hyperledger Umbra: Simulating Hyperledger Blockchains using Mininet - Raphael V. Rosa, UNICAMP

Description

Hyperledger Umbra: Simulating Hyperledger Blockchains using Mininet - Raphael V. Rosa, UNICAMP

The simulation research internship during the summer of 2018 led to the creation of the Hyperledger Umbra Lab. Due to the overall difficulty of getting Hyperledger blockchain frameworks running under the Shadow simulation tool, work on the Umbra lab has slowed to a crawl. A different network tool called Mininet has been proposed as an alternative to using Shadow and it has the potential to drastically reduce the startup cost of getting a network emulation tool running Hyperledger blockchains. Umbra stands for a research project to test the feasibility of running Hyperledger blockchains under Mininet for the purposes of running scalability and consensus mechanism experiments.

In 2019, Hyperledger Umbra was successfully implemented on Mininet, initially reproducing Fabric networks. Umbra can be extensible to emulate any Hyperledger project, and scalable to hundreds (even thousands - ongoing tests) of nodes.

The aim of showcasing Umbra, demonstrating a blockchain network easily deployed in a laptop, consists in building the foundations for designing and conducting empirical computer science research experiments to help direct improvements in design and architecture of future versions of the Hyperledger blockchain platforms. Examples of such improvements include new consensus algorithms, broadcast/gossip protocols, and security mechanisms.

In this technical presentation, Raphael V. Rosa is going to show Umbra, the platform architecture, its design principles and functionalities, how to create and run experiments on it, a Fabric network demonstration, and how to extend the platform to support other Hyperledger projects.

----
Hyperledger – Open Source Blockchain Technologies

Hyperledger is an open source community focused on developing a suite of stable frameworks, tools and libraries for enterprise-grade blockchain deployments.
It serves as a neutral home for various distributed ledger frameworks including Hyperledger Besu, Fabric, Sawtooth, Indy, as well as tools like Hyperledger Avalon, Cactus and libraries like Hyperledger Aries, Ursa. Learn more about Hyperledger projects: https://www.hyperledger.org/use

Case Studies: https://www.hyperledger.org/learn/case-studies
Training & Certification: https://www.hyperledger.org/learn/training
Tutorials: https://www.hyperledger.org/use/tutorials
Webinars: https://www.hyperledger.org/learn/webinars
Events: https://www.hyperledger.org/events
Vendor Directory: https://www.hyperledger.org/use/vendor-directory

Subscribe to the Hyperledger Newsletter: https://www.hyperledger.org/newsletter

Follow-us on Twitter @Hyperledger

Learn about Hyperledger Membership: https://www.hyperledger.org/about/join

#Hyperledger #HyperledgerGlobalForum2020 #Blockchain

A

So this project is called Umbra.

A

My name is Rafa I'm, a postdoc research University of Campinas in Brazil, and this project brought it's about how we can run hyper ledger blockchain projects in emulated environment.

A

It all started in 2018 when there was this internship project that it was a idealised to run the shadow simulation. That is mostly used by running a Bitcoin and the Tor network, and they saw that there are some issues with that because most of the libraries needed for the blockchain projects, we were not supported by shadow, so the project was basically solid and in 2019 I contacted David Hughes, be the mentor of that project and say: hey I'm.

A

Doing some tests here for a telco use case I'd like to see how it scales and I had this this platform called mini net. Let's use it for emulation in programmable networks and I think it can scale to what you need. Mini net is a basically a framework or a platform that is fast in simple to run program all networks in a laptop it was extended that it can runs in multiple servers to run containers and even abstractions of Wi-Fi.

A

So my Deva says: ok, since the high project projects provide container images as long as I support, this container images, I can run them in in this environment. The only thing I needed is to create the configuration to make it run and enable it. So the idea behind that design was that ok, I can specify a configuration file. That would say hey give me a something like a topology that I have some abstractions on on it.

A

This would generate an automated, a config file with all the crypto material, and that is going to create a scenario of that network, to which I could add smart contracts and so on. Events that would run the smart contract and a broker that would eat basically grab this sinner and instantiate it and play all these events at the topology. That's there and everything is going to be logged in and later use it for analysis.

A

So this was in the mind this was set up as the main goals of the the initial project is this internship project, so the basically the beauties happens, architecture evaluated this architecture against the the fabric project, build some stimulus, for example interacting with the fabric network, monitor it the containers and the links create some dynamics of the network like at the adding and removing links, shutting down and getting back nodes to into the network and at the end of the the demolition experiment.

A

That would build an analysis report that would give what happened in the post-mortem report so, and this would be extended for the repair, larger projects and Indian also been documented. So all of this was achieve it. Yet the only thing that missing is I could reproduce moot for other blockchain proteins in it. This is yet to finish, for example, I'm still coding the I row hop looking for for this I'm, going to tell you how it works.

A

Here's the basic architecture of this Umbra, so what it happens basically, is there's a design API. This blue component, the that defines some API, is that abstract? What would be a configuration file for that, so a user grabs? This API is composed the scenario that is basically set of topology and events and runs this scenario using this other component called Umbra broker the green one. The Umbra broker interacts with the other kampala G, with the other component called ingress scenario that deploys the topology and the actual infrastructure topology.

A

Basically, the containers which the blockchain nodes and the links interconnecting them and the lumber broker then realized the events in this this. This topology, for example, calling the chain codes to be vocal deployed and, and so on. So.

A

Basically, talking about the each one of these components in the uber design, there are API for multiple blockchain projects like that create multiple blockchain topologies in different aspects. I mean different nodes, interconnected with different abstractions.

A

These nodes can have restrictions of CPU memory disk and also the links and have restrictions of bandwidth, delay and packet loss. So, in the end, when you call these API is to build this scenario, they're going to generate all the the material needed to run. That scenario, for example, in the case of fabric, could be all the cryptokey, the certificates or the environment environmental variables in a single folder, this material generated by this number design. It is going to be placed it in the Umbra broker.

A

That is basically the main component that is going to consume. This scenario configuration topology and events, and this component is going to request the topology deployment into the other component hungry scenario. That instance she is the topology and later on, acknowledging that topology, the Umbra dorker broker basically interacts with the hyper ledger. Blockchain basically interacts with the infrastructure and also the the blockchain components. So, for example, we using as the case it can interact with each one of the nodes running the in the blockchain in this.

A

This is all takes place in a plugin oriented manner. It means that a different lock chain as the case and can be added to this this this component, so the the the the basically the architecture is fully modular, and this is the Umbra scenario. This is the one that makes use of this mini net, or here called container net API. That actually stands.

A

She ate the containers and the links tinker, connect them and configure them with the restrictions of CPU memory and bandwidth, packet loss and so on, and ideally this component can span across multiple hosts.

A

This is actually the the scaling test that I'm testing now that to make it run across different servers, so to scale to hundreds and thousands of nodes, and example that was coded in this project to showcase it that it's part of the demo is that I got the fabric project and version 1.4 I I saw the main documentation of it and got all the containers set.

A

So there is the requirements to run them created all this abstraction of this class called fabric topology that adds years orgs and order to treat, and it calls off all of these binary to generate the crypto material needed for each one of these containers fuels all of this. This darker component, there are containers descriptors with all the the proper paths volumes, environmental variables and port mappings. This is all done transparently to the user, to the tester, generating all this config files.

A

Finally, it generates just one single config file that reference is all of this configuration that is used by this broker component, that loads this scenario and instantiate it using the the scenario component and triggers this fabric events and this fabric events is basically all the events that are offered by the fabric, Python SDK, to interact with for the nodes, the CAS and the order.

A

So this was running successfully in a laptop easily, basically in a hypothetical way, if you like to say: hey I'd like to test a new contracting and lossy network with a particular graph abstraction I mean if your network is like a star or a different kind of interconnected the device you can build. Your this network by hand specifying like just a single config file. It's just a simple API, like add node, add link from this node to this node and so on.

A

There are different types of distractions that I'm planning to to implement here, like this, create a network with five orgs and three notes per organ and with a simple profile of CPU memory and disk, and so on.

A

If you'd like to run a specific topology graph for that I mean you'd like to see like how it scales to multiple servers and these servers you can place here how they're the interconnection of these containers happens inside the servers and also outside the service, and there's also demand that the possibility to monitor all of these.

A

These features in the inter components in the runtime, like the type the number of the lay of the messages in the consensus algorithm, for example, and in this way, if you think what this project actually enables is I mean, besides of being a generic platform for multiple hyper ledger block chains, what Dave, Hughes bead-like the mentor of this project was talking to me initially is like okay.

A

We now have some manners to make some irreducible academic work on this blockchain projects, so we have wasted the fine, comparable methodologies for consensus for gossipy protocols and even for security mechanisms.

A

The idea now is that we would like to test these at scale so initially when they talk to me and I say: okay, this runs now, let's put this in a in a cloud and make it run in a hundreds and thousands of containers and see how it works, how it behaves- and this is the task now and and after that we see that okay, it works with monitor everything. Everything is working. Fine. We are going to make some empirical computer science research experiments.

A

Basically, we are going to say hey if you are doing a new feature to this blockchain project. We have a perfect environment here for you to test with like few lines of code and that's all and the code first acknowledging that they've history is the the mentoring, the idea, at least after all, this project I also acknowledge the the internship community. This is a part of the internship last year and the code is fully available in the github. This is a part of a project part of hyper ledger labs.

A

This is just let's say a newborn project, so there's still a lot of source code to be optimized. It here the documentation, the basic documentation just to understand the architecture, the components and how they interact. They architecture fuel flow view directory structure. It's fully placed in the this read the docs files and I hope you enjoyed this project and I'm looking for ideas and how we can explore this project from now on. Thank you.

A

A

Now now I'm doing this because I'm maintained here, but there's a new project posted in the internship this year that they've posted and this project actually addressed the scalability of this. This project, I can extend it to the scalability trials.

A

But now I'm keeping I'll keep my D initially was to measure the scalability in telecom use case that I'm Marshall, I'm, actually part of networking field, so I was playing with that into the telecom use case. I grabbed this as an opportunity implemented. It saw it how it realizes they're, still monitoring features in the number that I did not push it to the mainstream code, but it'll be push it soon, and the idea of me is okay.

A

If you give me a news case, this platform should give you the footprint of that use case in different manners. That's the my view and I'm pushing that code online.

A

Thank you guys.