I’m fortunate to work with talented technical leaders who are helping to drive innovation forward in strategic areas of technology. My guest blogger this week is Frank Michaud, who is working with Cisco and the OpenFog Consortium to help develop a common industry approach to implementing blockchain in a highly distributed fog computing environment.
One of the challenges in a distributed computing environment such as fog computing is how to safeguard network resources and transactions with an equally distributed security structure. Fog computing creates a sort of mesh system where all nodes have equal roles, based on their capacities, and onto which we distribute computational loads. Along with distributed computing, we need distributed trust and security. This is especially important for cases where the infrastructure and layers of the fog node stack are owned and manage by many different entities.
How do you manage trust in a decentralized, distributed manner among actors that don’t necessarily trust each other? Blockchain technology is made for this kind of challenge.
There is a natural match between fog computing and blockchain—or distributed ledger technology. This is not only true in cases where different actors need to work together but don’t trust each other, it’s also true for disconnected or autonomous systems. The autonomous features of fog computing require trusted transactions in situations where a system has to function while disconnected from the cloud or data center. As fog technology becomes a critical part of more Internet of Things (IoT) and 5G applications, there is a growing need to leverage distributed ledger technology to establish consensus for each transaction.
Not every blockchain consensus mechanism is suitable for fog applications. For example, ”Proof-of-Work” (PoW) consensus requires massive computing capacity and power to solve a complex puzzle, so couldn’t be hosted on a fog device. But there are many other protocols such as “Proof of Stake” (PoS) that are capable of running on nodes with similar capacities of fog nodes.
As part of its mission to accelerate the adoption of fog computing, the OpenFog Consortium is working on building an interoperable and composable architecture for blockchain in fog environments. That implies having different entities in the system that don’t trust, or even know, each other but need to make a consensus decision in a distributed system. The Consortium’s work in this area supports “autonomy,” one of the eight pillars of the OpenFog Reference Architecture.
This is why we are already seeing several blockchain-oriented startups join the OpenFog Consortium, including:
- iExec: The first blockchain-based marketplace for decentralized cloud computing
- Hyperchain: Enterprise-level blockchain platform for fraud prevention, data trading, supply chain, and securities
- KeyChain: Global blockchain data security infrastructure for finance, industry and enterprise
- Xage: The first blockchain-protected security platform for Industrial IoT
- SONM: Distributed cloud services such as PaaS and IaaS, based on fog computing and secured by blockchain
- Aetherworks: Original software for distributed systems, including software-defined storage and fog computing
The systems for distributing trust among different actors in a fog environment must be rooted in the OpenFog architecture framework to assure interoperability. This will enable horizontal use cases where distributed system providers can commoditize and sell fog computing resources, with the assurance of being paid for these services. Any resource connected to an OpenFog system could then become a commodity of such a marketplace.
A group within the OpenFog Consortium is currently working on setting up the foundational requirements to enable distributed ledgers. This work will lead to additional requirements in the architecture framework, as well as the creation of a formal working group. Stay tuned to learn how these efforts develop.
I didn’t realize that fog gateways and edge nodes can be chained together to provide a better coverage and/or different layers of access and services. Wouldn’t purposely build edge or gateway devices provide the right consensus mechanism?