New trends and ideasA systematic study of double auction mechanisms in cloud computing
Introduction
Recently, a tremendous growth has been observed in the adoption of cloud computing by various organizations owing to various technological advancements and the appealing business centric nature of the cloud. Cloud computing is emerging as an important utility and is growing rapidly well supported by the major service providers e.g. Amazon, Microsoft, IBM, Google to name a few. The paradigm is successful with the establishment of various interoperability, security and QoS standards coming in place (Sajid and Raza, 2013). The main goal of cloud computing is to deliver storage, network, servers, computing or their combination as a service to the users (Buyya et al., 2013). NIST defines five basic characteristics of cloud computing as: on demand self-service, broad network access, resource pooling, elasticity and measured service (Snaith et al., 2011). On demand provisioning of services and pay per use model has become the main features offering big advantage of cloud computing (Garg et al., 2013).
The growing popularity of cloud computing services demands attention on the business aspects of the cloud for the resource provisioning and pricing issues. In a cloud market, various cloud providers want to sell their resources amongst multiple potential cloud users. The concept of the cloud market is still new and is different from other conventional markets in several aspects such as resource pricing mechanisms, service provisioning strategies, resource allocation rules and payment mechanisms. In a cloud market, a cloud provider competes with other resource providers with respect to expected pricing and QoS in order to maximize its Return on Investment (ROI). On the other hand, cloud users compete with each other for the required (limited) resources in least pricing and good QoS minimizing the total procurement cost. Thus, a cloud user is interested to benefit all the inherent features of the cloud viz. scalability, elasticity, flexibility, disaster recovery and at the same time, the cloud service providers vies to maximize their profit. A major challenge in such a cloud system thus becomes to find an optimal market based on unbiased strategy and offering the best possible satisfaction to both the cloud provider and the cloud user.
Game theory and mechanism design (Narahari, 2014) provides a basic framework to design and study such competitive markets. Also, the market strategy that is unbiased for both the user and the provider is highly desirable. As both the user and the provider are rational, intelligent and competing in the cloud market, auction based models are quite useful to model in such types of situation (Samimi et al., 2015). In auction, the price is determined by the supply and demand of the resources. Auctions are easy to implement, decentralized and suitable for distributed systems e.g. grid computing, cloud computing etc. (Buyya et al., 2002). It is also one of the many ways to implement the dynamic pricing. Dynamic pricing mechanism is desirable in cloud market because of its various advantages over the fixed pricing strategy. Dynamic pricing increases the total revenue of the resource provider by frequently changing its selling price which depends upon the supply/demand of the resources and many other factors (Narahari et al., 2005). It also creates the healthy competition among the users and increases the efficiency of the cloud resource usage (Mihailescu and Teo, 2010). A real example of such dynamic pricing implementation in cloud systems is Amazon's Spot market (“Amazon EC2 Spot Instances”) an auction based cloud market. Amazon sells its spare capacity as spot instances without disclosing any information about the auction mechanism and the spot price mechanism. In recent, Google has also initiated offering cloud services based on the dynamic pricing.
Mainly, there are three types of auction: Forward, Reverse and Double auction based on the bidding structure. In forward auction, multiple buyers compete with each other by bidding for the resources offered by a single cloud provider. Reverse auction occurs between a single cloud user and multiple cloud provider. The user first reveals its call for proposal (CFP) by quoting the required resources. Multiple cloud providers then compete with each other by bidding for the resources required by this single user. In double auction bidding is done by both the players of the market i.e. multiple buyers and the multiple sellers. Till now, in the real cloud market, only forward auction has been used by Amazon to sell the underutilized spot instances (“Amazon EC2 Spot Instances”). However, in literature, various models based on all types of auctions are reported. This work focuses only on double auction mechanisms in cloud computing.
Researchers and professionals have studied the double auction mechanisms for cloud computing, as summarized in this work, but till now no real implementation of double auction has been done in a real cloud market. Though, some laboratory e-market supports the double auction mechanism (Wellman et al., 2001) but the auction properties are not guaranteed. In future, it is quite likely that companies and business world will start offering cloud services and cloud users will avail these services. A good competition is expected between cloud users and service providers. Various benefits of double auctions include dynamic pricing, efficient resource allocation, supply and demand principles, less time consumption (Narahari et al., 2005).
To the author's best knowledge, a double auction model for cloud computing market that is truthful for all the participants i.e. users as well as providers has not been proposed till date. Most of the reported work ensures the truthfulness for either the cloud users or cloud providers. This work presents an exhaustive study of the basic double auction mechanisms, their properties and their applicability with a detailed discussion in context to the cloud computing. Salient features of the proposed work are as follows:
- •
Most of the double auction mechanisms, proposed for cloud computing in the past, are detailed and compared in terms of various basic auction properties.
- •
A general framework for double auction in cloud market is designed which gives a direction to the researchers and business professionals to design double auction mechanisms for cloud computing environments.
- •
A multi-unit double auction based mechanism (TMDA) is proposed for the cloud computing market. TMDA is individual rational, budget-balance along with being truthful for both the cloud users and the providers.
- •
Various challenges of double auction mechanisms are discussed setting a future direction to explore and extend the double auction mechanisms for the cloud systems.
The rest of the paper is as follows. A thorough survey on the double auction mechanisms in cloud computing has been done in Section 2. In Section 3, a framework for double auction in the cloud market is presented followed by a truthful multi-unit double auction (TMDA) model for the cloud resource allocation in Section 4. Section 5 discusses various issues and challenges for the double auction based cloud computing market. Section 6 concludes the work.
Section snippets
Double auction
McMillan (1994) documented the Federal Communications Commission (FCC) auction held in 1994 for selling the spectrum using auction which generated interest among researchers and scholars too working in the domain of auction. Another reason for bolstering the growth of auction based approaches is due to explosion of B2B exchanges. For example, a special application of auction is the procurement of complementary goods. Forrester has estimated that in 2003 (www.forrester.com), B2B trade value
A framework for double auction in cloud market
With the increasing popularity of cloud computing, the size of cloud market has also increased at a fast pace. At present, a number of cloud providers offer cloud services to the cloud users with differential pricing and varying QoS attributes. With time, more and more startups and big corporates are expected to adopt the cloud platform. In line, there are large number of cloud users who want to execute their tasks or jobs using cloud resources with minimum prices and better QoS. Cloud
Truthful multi-unit double auction (TMDA) in cloud computing
In previous section, numerous double auction mechanisms have been presented in cloud computing. Most of these mechanisms focus on the allocation function which matches the most eligible cloud users and providers for resource trading. A few of these works focus on the design of truthful payment schemes but the truthfulness is considered only for the user. As, both side bidding is involved, the auction mechanism should ensure to be truthful for both the sides in a sustainable cloud market. As per
Some issues and future directions
Some issues that arises from the study have been pointed out here along with some future research directions.
Discussion and conclusion
Market-based Double Auction mechanisms are poised to play a greater role to implement the dynamic pricing and have recently been infused for selling the underutilized and spare cloud resources by even dominant cloud providers such as Amazon. In a cloud market where VMs are traded among its users and providers, the double auction framework becomes very appropriate helping to avail the auction benefits for cloud services. This work discusses the basic double auction mechanisms in detail
Dinesh Kumar is Ph.D. student in computer science at the School of Computer and Systems Sciences, Jawaharlal Nehru University, New Delhi, India. His research interests include resource provisioning and pricing in cloud computing.
References (96)
- et al.
Quality of service approaches in cloud computing: a systematic mapping study
J. Syst. Softw.
(2015) - et al.
Incentive-compatible, budget-balanced, yet highly efficient auctions for supply chain formation
Decis. Support Syst.
(2005) - et al.
A fair multi-attribute combinatorial double auction model for resource allocation in cloud computing
J. Syst. Softw.
(2015) - et al.
Virtualization
Mastering Cloud Comput
(2013) - et al.
Price formation in double auction markets
J. Econ. Dyn. Control
(1996) - et al.
Trade reduction vs. multi-stage: a comparison of double auction design approaches
Eur. J. Oper. Res.
(2007) - et al.
An auction method for resource allocation in computational grids
Futur. Gener. Comput. Syst.
(2010) - et al.
Pricing in combinatorial double auction-based grid allocation model
J. China Univ. Posts Telecommun.
(2009) A dominant strategy double auction
J. Econ. Theory
(1992)- et al.
Efficient mechanisms for bilateral trading
J. Econ. Theory
(1983)
Bidding strategy for agents in multi-attribute combinatorial double auction
Expert Syst. Appl.
Multi-dimensional fairness for auction-based resource allocation
Knowl.-Based Syst
Multi-attribute auctions with different types of attributes: enacting properties in multi-attribute auctions
Expert Syst. Appl.
The multiple unit double auction
J. Econ. Behav. Organ.
Algorithm for optimal winner determination in combinatorial auctions
Artif. Intell.
BOB: improved winner determination in combinatorial auctions and generalizations
Artif. Intell.
Credibility-based cloud media resource allocation algorithm
J. Netw. Comput. Appl.
A test suite for the evaluation of mixed multi-unit combinatorial auctions
J. Algorithms
Solving the combinatorial double auction problem
Eur. J. Oper. Res.
Combinatorial auction-based allocation of virtual machine instances in clouds
J. Parallel Distrib. Comput.
Amazon Elastic Compute Cloud (Amazon EC2) [WWW Document]
Amazon EC2 Spot Instances [WWW Document]
Concurrent auctions across the supply chain
J. Artif. Intell. Res.
A cloud service selection model using improved ranked voting method
Concurr. Comput. Pract. Exp.
The Role of Reverse Auctions in Strategic Sourcing
Blueprint for the intercloud - protocols and formats for cloud computing interoperability
Competitive market Institutions: double auctions vs. sealed bid-offer auctions
Am. Econ. Rev.
Economic models for resource management and scheduling in grid computing
Concurr. Comput. Pract. Exp.
CloudSim: a toolkit for modeling and simulation of cloud computing environments and evaluation of resource provisioning algorithms
Softw.
Towards efficient Greedy allocation schemes for double-sided cloud markets
Truthful bundle/multiunit double auctions
Manage. Sci.
Truthful double auction mechanisms
Oper. Res.
Agent competition double-auction mechanism
Manage. Sci.
Multipart pricing of public goods
Public Choice
Cloud4SOA [WWW Document]
ECU [WWW Document]
Efficient Nash equilibrium based cloud resource allocation by using a continuous double auction
Market-Driven Continuous Double Auction Method for Service Allocation in Cloud Computing
The double auction market institution: a survey
Taming the computational complexity of combinatorial auctions: optimal and approximate approaches
IJCAI Int. J. Conf. Artif. Intell.
Applying double-sided combinational auctions to resource allocation in cloud computing
Mandi: a market exchange for trading utility and cloud computing services
J. Supercomput.
Incentives in teams
Econometrica
Inter-Cloud architectures and application brokering: taxonomy and survey
Softw.
Solving combinatorial auctions using stochastic search
An efficient method to find approximate solutions for combinatorial double auctions
Scalable multi-agent learning algorithms to determine winners in combinatorial double auctions
Appl. Intell.
Cited by (58)
A truthful double auction framework for security-driven and deadline-aware task offloading in fog-cloud environment
2024, Computer CommunicationsA secure and privacy-preserving protocol for holding double auctions in smart grid
2021, Information SciencesCitation Excerpt :These different mechanisms can provide us with various desirable properties such as economic efficiency, individual rationality, budget balance, computational efficiency, truthfulness, and fairness. Based on the environments in which the double auction is implemented, the double auction mechanism should satisfy different properties [30]. These mechanisms may differ in the way of determining the trading price or determining participants’ volume allocations.
Cyber physical system-enabled on-demand logistics trading
2021, International Journal of Production Economics
Dinesh Kumar is Ph.D. student in computer science at the School of Computer and Systems Sciences, Jawaharlal Nehru University, New Delhi, India. His research interests include resource provisioning and pricing in cloud computing.
Gaurav Baranwal is a faculty in Department of Computer Science and Engineering, MMMUT, Gorakhpur, UP, India. His research interests include resource provisioning and service coordination in cloud computing.
Zahid Raza is a faculty in the School of Computer and Systems Sciences, Jawaharlal Nehru University, India. He is M.Sc. in Electronics, M. Tech in Computer Science. He did his Ph.D. in computer science from Jawaharlal Nehru University, India. Prior to joining JNU, he served as a lecturer in Banasthali Vidyapith University, Rajasthan, India. His research interests include parallel and distributed systems, evolutionary algorithms and multi-objective evolutionary algorithms. He has published many peer-reviewed articles and has proposed various scheduling models for computational grid, cloud and cluster systems.
Deo Prakash Vidyarthi is a professor in the School of Computer and Systems Sciences, Jawaharlal Nehru University, New Delhi. He was associated with the Department of Computer Science of Banaras Hindu University, Varanasi for more than 12 years before joining JNU as an associate professor. Dr. Vidyarthi has published around 75 research papers in various peer reviewed International Journals and Transactions (including IEEE, Elsevier, Springer, Wiley, World Scientific etc.) and around 45 research papers in proceedings of various peer-reviewed conferences in India and abroad. Dr. Vidyarthi has authored two books. One entitled “Technologies and Protocols for the Future Internet Design: Reinventing the Web” published by IGI-Global (USA) released in Feb. 2012, and another entitled “Scheduling in Distributed Computing Systems: Design, Analysis and Models” published by Springer, USA released in 2009. He also has contributed chapters in many edited books. He is in the editorial board and in the reviewer's panel of many International Journals. Dr. Vidyarthi is the member of the IEEE, senior member of the International Association of Computer Science and Information Technology (IACSIT), Singapore, International Society of Research in Science and Technology (ISRST), USA and International Association of Engineers. Research interest includes Parallel and Distributed System, Grid and Cloud Computing, Mobile Computing and Evolutionary Computing.