CONF contributions to the traditional cloud industry

“At the heart of the cloud strategy is a cloud-first approach with a secure, hybrid, multi-cloud service offering.”

European Commission Cloud Strategy (source)

Our last blog post discussed preliminary ideas of building a more open, fair, and trustworthy decentralised cloud marketplace for all service providers and customers. In this current post, we discuss the challenges faced by a decentralised cloud marketplace and our progress in designing and implementing the ARTICONF ecosystem using blockchain technologies.

Decentralised Cloud Marketplace Requirements and Challenges

Cloud services play an increasingly important role in industrial innovation and scientific applications by processing information in real-time running live simulations and enabling extensive collaborations. Several industrial use cases need to employ multiple data centres or service providers to handle decentralised collaboration between data, resources, and customers. We highlight three main requirements and challenges in this context:

Provider selection, service customisation, and capacity planning.

The current cloud service model typically adopts a vendor lock-in model. ​​In a decentralised cloud marketplace, however, customers need to select cloud services from different providers from distributed locations (e.g. sensors or repositories) with diverse data access (e.g. medical) and performance constraints (e.g. real-time decisions, early warnings). Cloud service providers’ various price and reputation models make the selection process time-consuming and difficult to optimise. 

SLA interoperability and guarantee.

Time-critical applications (e.g. crowd media processing news) require thorough optimisation of their logic and components to guarantee the cloud service quality (i.e. virtual machines, network connectivity). The diverse SLA terms among providers and the uncertainties in the SLA guarantees make performance optimisation difficult. 

Virtual infrastructure automation.

Applications involving multiple providers or data centres require automation of the virtual infrastructure provisioning, software platform, application components deployment (often in isolated containers), application monitoring, and adaptation. However, the heterogeneity in application programming interfaces from different providers raises interoperability issues across those providers. This lack of interoperability makes automated provisioning and deployment a challenge, resulting in high complexity for monitoring the runtime infrastructure quality and detecting SLA violations for infrastructural adaptation.

How to Build an Auction and Witness Enhanced Decentralised Cloud Marketplace?

ARTICONF is a recent EU project that aims to develop tools for decentralised social media applications. One of the open-source tools is CONF for planning cost-effective distributed cloud infrastructures in a trustworthy way. In CONF, a customisable auction model selects the most suitable cloud customers/providers. A new role called “market witness” introduced in the traditional cloud service delivers complete lifecycle monitoring of platform performance. The role of the witness is an anonymous participant in the blockchain system who desires to gain rewards by reporting violations. The incentive mechanism for different witnesses in CONF ensures that witnesses always behave honestly to maximise profit. In this way, a new blockchain-based decentralised cloud market model tackles the challenge of detecting SLA violations responsibly. 

To meet the requirements of a decentralised cloud marketplace, we designed three smart contracts using the Ethereum blockchain (i.e. auction contract, witness contract, and SLA contract) to support trustworthy and fair interactions between different cloud stakeholders. The interaction between the contract factory and the various subcontracts is as follows. First, a market manager calls the contract factory to create a new auction contract. Then, an auction contract with customised auction rules for the business requirements supports transparency and automation of the auction process. In this case, decentralised service providers can register and submit their bids for services using commitments on the blockchain.

A commitment scheme is a cryptographic primitive that allows a person to commit to a selected value and reveal it later. The use of commitment schemes can protect the privacy of auction participants well. The auction contract then chooses the winning providers based on the highest k bids and generates k SLA contracts for each provider. After settling the auction (and still delivering the services), the market manager calls the contract factory again to create a witness contract with customised incentive mechanisms to encourage truth-telling witnesses. Our previous research [2] mentions more details about our game theory-based witness payoff design for cloud SLAs. Then, different winner providers deploy cloud services off-chain while the witnesses monitor the services. If the QoS satisfies the requirements in the SLA contract, there is no violation. Otherwise, there is a violation. The service monitoring results are also returned to the auction contract and determine the status of the auction.

Regarding the implementation details of these smart contracts, we first designed some function interfaces for each stakeholder. These function interfaces require certain computational tasks on-chain and therefore require specific transaction fees. Nevertheless, we find that the transaction involved in the three contracts is still very economical compared to the enormous commission fee in the case of traditional auction houses that do not rely on the blockchain. 

In summary, CONF aims to improve the existing infrastructure support in the ARTICONF tools for social network applications to reach the following objectives: 

  1. Improve the efficiency of selecting service providers from a decentralised marketplace by proposing a decentralised auction framework; 
  2. Enhance the trustworthiness of SLA among providers and users by using a dynamic decentralised witness model;
  3. Improve verification efficiency of smart contracts in the cloud service marketplace by using an interactive business simulator;
  4. Improve the continuous DevOps efficiency of the decentralised marketplace by integrating our tool with the DevOps framework. Some subsystem features are still under development and are part of our future work. 

We plan to develop a highly modular software architecture for a decentralised cloud ecosystem in the CONF tool. Some subsystem features (e.g., the cross-chain simulator in blockchain planner) are still under development. In the future, we will continue to test our framework and demonstrate its feasibility in the EU ARTICONF project.

References

  • [1] Shi, Z., Farshidi, S., Zhou, H., & Zhao, Z. (2021, September). An Auction and Witness Enhanced Trustworthy SLA Model for Decentralized Cloud Marketplaces. In Proceedings of the Conference on Information Technology for Social Good (pp. 109-114).
  • [2] Zhou, H., Ouyang, X., Su, J., de Laat, C. and Zhao, Z., 2019. Enforcing trustworthy cloud SLA with witnesses: A game theory-based model using smart contracts. Concurrency and Computation: Practice and Experience, p.e5511.
  • [3] Shi, Z., Zhou, H., Surbiryala, J., Hu, Y., de Laat, C. and Zhao, Z., 2019, December. An Automated Customization and Performance Profiling Framework for Permissioned Blockchains in a Virtualized Environment. In 2019 IEEE International Conference on Cloud Computing Technology and Science (CloudCom) (pp. 404-410). IEEE Computer Society.
  • [4] Uriarte, R.B., Zhou, H., Kritikos, K., Shi, Z., Zhao, Z. and De Nicola, R., 2020. Distributed service‐level agreement management with smart contracts and blockchain. Concurrency and Computation: Practice and Experience, p.e5800.

This blog post was written by The University of Amsterdam team.

< Thanks for reading. We are curious to hear from you. Get in touch with us and let us know what you think. >