Roman Matzutt, Oliver Hohlfeld, Martin Henze, Robin Rawiel · 6 authors
As transaction fees skyrocket today, blockchains become increasingly expensive, hurting their adoption in broader applications. This work tackles the saving of transaction fees for economic blockchain applications. The key insight is that other than the existing "default'' mode to execute application logic fully on-chain, i.e., in smart contracts, and in fine granularity, i.e., user request per transaction, there are alternative execution modes with advantages in cost-effectiveness. On Ethereum, we propose a holistic middleware platform supporting flexible and secure transaction executions, including off-chain states and batching of user requests. Furthermore, we propose control-plane schemes to adapt the execution mode to the current workload for optimal runtime cost. We present a case study on the institutional accounts (e.g., coinbase.com) intensively sending Ether on Ethereum blockchains. By collecting real-life transactions, we construct workload benchmarks and show that our work saves 18%\sim 47%18%-47% per invocation than the default baseline while introducing 1.81%\sim 16.59%1.81%-16.59% blocks delay.
Ikuo Magaki, Moein Khazraee, Luis Vega Gutierrez, Michael Taylor
GPU and FPGA-based clouds have already demonstrated the promise of accelerating computing-intensive workloads with greatly improved power and performance. In this paper, we examine the design of ASIC Clouds, which are purpose-built datacenters comprised of large arrays of ASIC accelerators, whose purpose is to optimize the total cost of ownership (TCO) of large, high-volume chronic computations, which are becoming increasingly common as more and more services are built around the Cloud model. On the surface, the creation of ASIC clouds may seem highlyimprobable due to high NREs and the inflexibility of ASICs. Surprisingly, however, large-scale ASIC Clouds have already been deployed by a large number of commercial entities, to implement the distributed Bitcoin cryptocurrency system. We begin with a case study of Bitcoin mining ASIC Clouds, which are perhaps the largest ASIC Clouds to date. From there, we design three more ASIC Clouds, including a YouTube-style video transcoding ASIC Cloud, a Litecoin ASIC Cloud, and a Convolutional Neural Network ASIC Cloud and show 2-3 orders of magnitude better TCO versus CPU and GPU. Among our contributions, we present a methodology that given an accelerator design, derives Pareto-optimal ASIC Cloud Servers, by extracting data from place-and-routed circuits and computational fluid dynamic simulations, and then employing clever but brute-force search to find the best jointly-optimized ASIC, DRAM subsystem, motherboard, power delivery system, cooling system, operating voltage, and case design. Moreover, we show how data center parameters determine which of the many Pareto-optimal points is TCO-optimal. Finally we examine when it makes sense to build an ASIC Cloud, and examine the impact of ASIC NRE.
In Bitcoin system, transactions are prioritized according to attributes such as the remittance amount and transaction fees, and transactions with low priority are likely to wait for confirmation. Because the demand of micro payment in Bitcoin is expected to increase due to low remittance cost, it is important to quantitatively investigate how the priority mechanism of Bitcoin affects the transaction-confirmation time. In this paper, we analyze the transaction-confirmation time by queueing theory. We model the transaction priority mechanism of Bitcoin as a priority queueing system with batch service, deriving the mean transaction-confirmation time. Numerical examples show how the demand of transactions of low remittance amount affects the transaction-confirmation time. We also consider the effect of the maximum block size on the transaction-confirmation time.
Alexander Chepurnoy, Mario Larangeira, Alexander Ojiganov
Bitcoin is the first successful decentralized global digital cash system. Its mining process requires intense computational resources, therefore its usefulness remains a disputable topic. We aim to solve three problems with Bitcoin and other blockchain systems of today by repurposing their work. First, space to store a blockchain is growing linearly with number of transactions. Second, a honest node is forced to be irrational regarding storing full blocks by a way implementations are done. Third, a trustless bootstrapping process for a new node involves downloading and processing all the transactions ever written into a blockchain. In this paper we present a new consensus protocol for Bitcoin-like peer-to-peer systems where a right to generate a block is given to a party providing non-interactive proofs of storing a subset of the past state snapshots. Unlike the blockchain systems in use today, a network using our protocol is safe if the nodes prune full blocks not needed for mining. We extend the GKL model to describe our Proof-of-Work scheme and a transactional model modifications needed for it. We provide a detailed analysis of our protocol and proofs of its security.
We need to shift our perspective of blockchain, from just the programmable ledger, to a networked infrastructure of computing machinery. Doing so, we could easily visualize how computer programs will operate over this new infrastructure, which is presently being used for transaction validations. We cannot take the cloud computing analogy literally because the blockchain infrastructure canât replace cloud computing completely. Although, it unbundles and democratizes cloud computing. Thus, companies and individual consumers can soon keep their data in distributed cloud network based on highly secure blockchain technology.
A purely anti-node non-peer-to-peer version of electronic money would allow online payments to be sent directly from one person to another without going through a financial institution and node promoters such as digital and gold-miners. Digital signatures provide part of the solution, but the main benefits are lost if a trusted third party is still required to prevent double-spending. We propose a solution to the double-spending problem using an anti-node non-peer-to-peer network which is not only distributed but encrypted to the core of the gene. The network timestamps transactions by hashing them into an ongoing chain of hash-based genesis-concept combined with military grade chain-reinforced-encryption using AES, SHA, RSA and custom algos, forming a record that cannot be changed without redoing the genesis-concept. The longest chain not only serves as proof of the sequence of events witnessed, but proof that it came from the largest pool of CPU power. As long as a majority of CPU power is controlled by nodes that are not cooperating to attack the network, they'll generate the longest chain and outpace attackers. The network itself requires minimal structure. Messages are broadcasted on a best effort basis, and nodes can leave and re-join the network at will, accepting the longest proof-of-work chain as proof of what happened while they were gone. However, in the proposed model, the central authority server itself acts as the node and represents the network transport layer by itself which enables it to not to be dependent on individual nodes, and the crypto-formulation continuous at server level. By this process, the dependency on nodes are ruled out and the server (the executioner, in this case) gets all the bits combined in one platform from the beginning to the end using enhanced salting layer, providing the facility of de-centralized defacto standard e-payments within nano seconds or less i.e., account to account transfer using one central block processing schema.
The growing parallelism and heterogeneity of modern computing infrastructures such as High Performance Computing (HPC) platforms raises new challenges to their programmers and users. Additional requirements have emerged nowadays, such as minimizing the consumed energy, reducing the utilized system resources, or providing built-in reliability mechanisms. Therefore High Performance Computing (HPC) applications require adaptation mechanisms and then must avoid traditional monolithic centralized approaches in favor of novel autonomous, flexible and decentralized decision systems. In this context, we describe here a dynamic and flexible adaptation scheme based on a Multi-Agent System (MAS) to handle parallel or distributed executions in an HPC environment. More precisely, we model and extend the existing HPC middleware Kaapi to offer the power of the ParaMoise multi-agent organizational framework. Our proposed solution, named ParaMASK, relies on the similarities between ParaMoise workflow-based functional specifications and the Direct Acyclic Graph (DAG) representation of the distributed execution within Kaapi. As a result, ParaMASK permits to analyze and reorganize the scheduling of tasks that compose a program in an autonomous and decentralized way, while additionally handling dynamic adaptations (using task migration to fulfill energy consumption goals for example). The proposed solution was implemented on top of the existing Kaapi middleware and includes an optimized algorithm for the agent coordination. ParaMASK has been validated with a series of experiments on a real computational grid. Experimental results show a good scalability and an exceptional low overhead induced by the approach: less than 1.5% execution time increase with periodic coordinations every 15 seconds on 2662 cores.
Grid computing is the computing paradigm that is concerned with coordinated resource sharing and problem solving in dynamic, autonomous multi-institutional virtual organizations. Data exchange and service allocation between virtual organizations are challenging problems in the field of Grid computing, due to the decentralization of Grid systems. The resource management in a Grid system ensures efficiency and usability. The required efficiency and usability of Grid systems can be achieved by building a decentralized multi-virtual Grid system. In this thesis we present a decentralized multi-virtual resource management framework in which the system is divided into virtual organizations, each controlled by a broker. An overlay network of brokers is responsible for global resource management and managing the allocation of services. We address two main issues for both local and global resource management: 1) decentralized allocation of tasks to suitable nodes to achieve both local and global load balancing; and 2) handling of both regular and broker failures. Experimental results verify that the system achieves dependable performance with various loads of services and broker failures.
Cloud computing is an emerging evolutionary computing model that provides highly scalable services over high-speed Internet on a pay-as-usage model. However, cloud-based solutions still have not been widely deployed in some sensitive areas, such as banking and healthcare. The lack of widespread development is related to usersâ concern that their confidential data or privacy would leak out in the cloudâs outsourced environment. To address this problem, we propose a novel active data-centric framework to ultimately improve the transparency and accountability of actual usage of the usersâ data in cloud. Our data-centric framework emphasizes âactiveâ feature which packages the raw data with active properties that enforce data usage with active defending and protection capability. To achieve the active scheme, we devise the Triggerable Data File Structure (TDFS). Moreover, we employ the zero-knowledge proof scheme to verify the requestâs identification without revealing any vital information. Our experimental outcomes demonstrate the efficiency, dependability, and scalability of our framework.
Welcome to the special issue of Concurrency and Computation: Practice and Experience (CCPE) journal. This special issue compiles a number of excellent technical contributions that significantly advance the state-of-the-art in autonomic cloud computing. Cloud computing 1, 2 is an emerging utility computing model that allows users to dynamically access, select, and configure a large pool of IT resources (virtual machine templates, storage, and networking elements) and deliver them as âcomputing utilitiesâ to consumers in a pay-as-you-go manner. Several vendors have emerged in this space including IBM, VMware, Microsoft, Manjrasoft, and Yahoo. This model of computing is quite attractive, especially for small and medium sized enterprises, as it allows them to focus on consuming or offering services on top of cloud infrastructure. At high-level, cloud computing might not seem radically different from the existing paradigms: World Wide Web, grid computing, and cluster computing. However, key differentiators of cloud computing are its technical characteristics such as on-demand resource pooling or rapid elasticity, self-service, almost infinite scalability, end-to-end virtualization support, and robust support of resource usage metering and billing. Additionally, nontechnical differentiators include services that are offered under pay-as-you-go-model, guaranteed Service Level Agreement (SLA), faster time to deployments, lower upfront costs, little or no maintenance overhead, and environment friendliness. Unpredictability is a fact in a distributed computing environment, and the Cloud is no exception. Performance unpredictability 3 in the Cloud is in fact a major issue for many users and it is coined as one of the major obstacles for cloud computing. For instance, researchers (biologists, physicists, finance analysts, etc.) expect guaranteed performance for their experiments, independent of the current workload and state 4 of IT resources of the Cloud, because this is key to repeatability of results. Other examples are small and medium sized enterprises (gaming company, web application providers) that want strict assurance on SLA; for example, an end-user request for a web page or multimedia content has to be served within the agreed time-limit. Hence, it is highly important for Cloud vendors that they have the ability to offer guaranteed SLAs based on performance metrics â such as response time and throughput. Interestingly, vendors seem to base their SLAs on availability of their offering, while completely ignoring response time and throughput. Hence, it is clear that dealing with performance unpredictability is critical to exploiting the full potential of clouds. In this special issue, we have tried to compile some high quality papers that exhaustively deal with some of the aforementioned issues. Next, we briefly describe the technical contributions, which were selected for publication in this special issue. All of the selected papers underwent a rigorous peer-review process. The end-to-end QoS negotiation for SLA establishment for composite services involves compound multiparty negotiations in which the composite service provider concurrently negotiates with multiple candidates for each atomic service, selecting the one that best satisfies the atomic service QoS preferences while ensuring that the end-to-end QoS requirements are also fulfilled. It is necessary to derive the atomic utility boundaries from the global utility boundary to be able to negotiate with potential candidates. Additionally, there has to be a mechanism for updating these boundaries in subsequent negotiation rounds based upon the individual negotiation outcomes. To counter these complexities, in paper 5 titled âEstablishing Composite SLAs through Concurrent QoS Negotiation with Surplus Redistributionâ, Richter et al. propose an algorithm for the decomposition of global utility boundary into atomic service utility boundaries, and the surplus redistribution from successful negotiation outcomes among the remaining negotiations. The proposed mechanism is a practical approach to efficiently coordinate concurrent service negotiations within complex workflows, enabling the iterative and interactive adjustment of the negotiation boundaries for each atomic service in a composition based on the performance of other atomic negotiations. They demonstrate the feasibility of our approach by evaluating it with some popular negotiation strategies using the Specialised Property Search Scenario. Many scientific workflows are data intensive where large volumes of intermediate data are generated during their execution. Some valuable intermediate data need to be stored for sharing or reuse. Traditionally, they are selectively stored according to the system storage capacity determined manually. As doing science in the Cloud has become popular nowadays, more intermediate data can be stored in scientific cloud workflows based on a pay-for-use model. In the paper in 6 titled âA data dependency based strategy for intermediate data storage in scientific cloud workflow systemsâ, Yuan et al. build an intermediate data dependency graph (IDG) from the data provenance in scientific workflows. With the IDG, deleted intermediate data can be regenerated, and as such they develop a novel intermediate data storage strategy that can reduce the cost of scientific cloud workflow systems by automatically storing appropriate intermediate data sets with one Cloud service provider. The strategy has significant research merits, that is, it achieves a cost-effective trade-off of computation cost and storage cost and is not strongly impacted by the forecasting inaccuracy of data setsâ usages. Meanwhile, the strategy also takes the usersâ tolerance of data accessing delay into consideration. Authors utilize Amazon's cost model and apply the strategy to general random and specific astrophysics pulsar searching scientific workflows for evaluation. The results show that our strategy can reduce the overall cost of scientific cloud workflow execution significantly. Recall that, one of the biggest premises of cloud computing is the flexibility of delivering IT resources and virtual appliances as an utility such as phone, electricity, gas, and water services. It enables users to have access to computing infrastructure, platform, and software as services over the Internet. To be competitive, however, Cloud providers need to be able to adapt to the dynamic loads from users, not only optimizing the local usage and costs but also engaging into agreements with other clouds to complement local capacity. The infrastructure in which competing clouds are able to cooperate to maximize their benefits is called a Federated Cloud. Just as clouds enable users to cope with unexpected demand loads, a Federated Cloud will enable individual clouds to cope with unforeseen variations of demand. The definition of the mechanism to ensure mutual benefits for the individual clouds composing the federation, however, is one of its main challenges. Gomes et al. in their paper 7 âPure exchange markets for resource sharing in federated cloudsâ propose and investigate the application of market-oriented mechanisms based on the General Equilibrium Theory of Microeconomics to coordinate the sharing of resources between the clouds in a Federated Cloud. Several research institutions and universities own computational capacity that is not effectively utilized, thereby providing an opportunity for such institutions to use such capacity to offer Cloud services (to both internal and external users). However, the unreliability and unpredictability of these resources mean that their use in the context of an SLA is high risk, leading to a reduction in reputation and economic penalties in case of SLA violation. To overcome these challenges, in the paper 8 titled âTowards autonomic management for Cloud services based upon volunteered resourcesâ, Caton and Rana propose a methodology that addresses the issues of unreliability and unpredictability such that Cloud software services could be hosted upon volunteered resources. To enable the harnessing of these resources, they rely on autonomic fault management techniques that allow such systems to independently adapt to the resources they use based upon their perception of individual resource reliability. Using the proposed approach they were able to scale out the backend infrastructure of the Cloud service elastically (minimum 30 s per worker), opportunistically, and autonomically. To summarize, the authors address two key questions in their paper: Can a campus volunteer infrastructure be used in Cloud provisioning? and What measures are necessary to ensure reliability at the resource level? To improve the hosting and delivery of applications through cloud-based IT resources, Champrasert et al. in the paper 9 titled âExploring self-optimization and self-stabilization properties in bio-inspired autonomic cloud computingâ, describe architecture to build self-optimizable and self-stabilizable applications. The design of the proposed architecture, SymbioticSphere, is inspired by key biological principles such as decentralization, evolution, and symbiosis. In SymbioticSphere, each cloud application consists of application services and middleware platforms. Each service and platform is designed as a biological entity, and implements biological behaviors such as energy exchange, migration, reproduction, and death. Each service/platform possesses behavior policies, as genes, each of which defines when and how to invoke a particular behavior. SymbioticSphere allows services and platforms to autonomously adapt to dynamic network conditions by optimizing their behavior policies with a multi-objective genetic algorithm. Moreover, SymbioticSphere allows services and platforms to autonomously seek stable adaptation decisions as equilibria (or symbiosis) between them with a game theoretic algorithm. This symbiosis augments evolutionary optimization to expedite the adaptation of agents and platforms. It also contributes to stable performance that contains a very limited amount of fluctuations. Simulation results demonstrate that agents and platforms successfully attain self-optimization and self-stabilization properties in their adaptation processes. We hope that the readers will find the articles of this special issue to be informative and useful.
Computational grids enable the sharing, aggregation, and selection of (geographically distributed) computational resources and can be used for solving large scale and data intensive computing applications. Computational grids are an appealing target application for market-based resource allocation especially given the attention in recent years to âvirtual organizations â and policy requirements. In this paper, we present a framework for truthful, decentralized, dynamic auctions in computational grids. Rather than a fullyspecified auction, we propose an open, extensible framework that is sufficient to promote simple, truthful bidding by endusers while supporting distributed and autonomous control by resource owners. Our auction framework incorporates resource prediction in enabling an expressive language for end-users, and highlights the role of infrastructure in enforcing rules that balance the goal of simplicity for end users with autonomy for resource owners. The technical analysis leverages simplifying assumptions of âuniform failureâ and âthreshold-reliabilityâ beliefs.