PDF Publication Title:
Text from PDF Page: 013
Chapter 1 Introduction We depend upon computer systems in every aspect of life. We expect systems to respond quickly, to behave as expected, and to be available when needed. However, the components which make up these systems, such as computers and the networks which connect them, are not reliable. Distributed consensus is the problem of how to reliably reach agreement in the face of failures and asynchrony. This longstanding challenge is fundamental to distributed systems and, when solved, gives us the power to construct reliable distributed systems from unreliable components. Lamport’s Paxos algorithm [Lam98] has been synonymous with distributed consensus for two decades [Mal]. It is widely deployed in production, and has been the subject of extensive research to optimise, extend and better understand the algorithm. Despite its popularity, Paxos performs poorly in practice, its inflexible approach is heavyweight, unscalable and can be unavailable in the face of asynchrony and failures. This thesis re-examines the problem of distributed consensus and how we approach it. Firstly, we prove that Paxos is, in fact, one point on a broad spectrum of approaches to solving distributed consensus, opening the door to a new breed of performant, scalable and resilient consensus algorithms. Then, we explore some of the new algorithms made possible by this result; some of which are even able to achieve consensus where it was previously thought impossible. In the next section, we describe the de facto approach to consensus in modern distributed systems (§1.1). For readers who are unfamiliar with the field, we then outline the historical context of this research, focusing on how early formulations of the problem of distributed consensus shaped (and arguably limited) how it is solved today (§1.2). This is followed by our critiques of this widely adopted approach and thus our motivation for re-examining how we can solve consensus (§1.3). Next, we describe the methodology we chose for the investigation into consensus (§1.4) and highlight the surprising contributions made to the field of consensus as a result (§1.5). 13PDF Image | Distributed consensus
PDF Search Title:
Distributed consensusOriginal File Name Searched:
UCAM-CL-TR-935.pdfDIY PDF Search: Google It | Yahoo | Bing
Cruise Ship Reviews | Luxury Resort | Jet | Yacht | and Travel Tech More Info
Cruising Review Topics and Articles More Info
Software based on Filemaker for the travel industry More Info
The Burgenstock Resort: Reviews on CruisingReview website... More Info
Resort Reviews: World Class resorts... More Info
The Riffelalp Resort: Reviews on CruisingReview website... More Info
| CONTACT TEL: 608-238-6001 Email: greg@cruisingreview.com | RSS | AMP |