On the Feasibility of Bilaterally Agreed Accounting of Resource Consumption

Abstract

The services offered by Internet Data Centers involve the provision of storage, bandwidth and computational resources. A common business model is to charge consumers on a pay-per-use basis where they periodically pay for the resources they have consumed (as opposed to a fixed charge for service provision). The pay-per-use model raises the question of how to measure resource consumption. Currently, a widely used accounting mechanism is provider-side accounting where the provider unilaterally measures the consumer's resource consumption and presents the latter with a bill. A serious limitation of this approach is that it does not offer the consumer sufficient means of performing reasonableness checks to verify that the provider is not accidentally or maliciously overcharging. To address the problem the paper develops bilateral accounting models where both consumer and provider independently measure resource consumption, verify the equity of the accounting process and try to resolve potential conflicts emerging from the independently produced results. The paper discusses the technical issues involved in bilateral accounting.

A Peer to Peer Protocol for Online Dispute Resolution over Storage Consumption

“…is unmonitorable, monitorable by a trusted third party, monitorable by one party, monitorable by both parties. In [5] the authors explain that the monitorability of a given parameter depends on several factors; among the most important ones are the accuracy of the metering, the physical location of the application (one or many) that affects the parameter, the physical location of the metering service and the trust assumptions about the metering service and its location.…”

“…Currently, most providers use unilateral provider-side accounting based on metrics collected by devices deployed within the providers' premises. An alternative and innovative approach is bilateral accounting where both the consumer and provider independently measure resource consumption and verify the parity of the accounting results [5]. A potential problem here is the emergence of potential conflicts derived from divergences between the independently produced accounting results.…”

Consumer Side Resource Accounting in the Cloud

“…Cloud service providers also perform their own measurements to collect usage data, although, as yet there are no equivalent facilities of consumer-trusted metering; rather, consumers have no choice but to take whatever usage data that is made available by the provider as trustworthy. A good introduction into the underlying trust issues can be found in [12].…”

CFaaS: bilaterally agreed evidence collection

“…The components could be located either within or outside the party's environment. In the latter case, the component's owner may need to take additional measures, such as the use of tamper-resistant mechanisms, to protect the component and its outputs against modification by other parties [68,69]. There are two approaches to producing a mutually trusted output: (i) A Trusted Third Party (TTP) produces the output using its own certified infrastructure, or (ii) the parties concerned use their respective unilaterally trusted outputs as the basis for agreement on a valid, mutually trusted output [70].…”

“…Here, the CCRP is a peer to peer dispute resolution protocol in a sense that it will be executed between the CSC and CSP without the intervention of a third party. The model utilizes a non-repudiable object sharing middleware for sharing the forensic data among CSC and CSP to form the basis of CCRP [68,[71][72][73]. The middleware can provide multi-party, non-repudiable agreement to a shared forensic data with the CSC and CSP holding their own copies.…”