Clement Research Statement
نویسنده
چکیده
Computer systems are a pervasive aspect of daily life. From the smart phones we use to access our favorite websites, to the services those websites provide, to the computer systems that manage core infrastructure services (e.g., air traffic control, the power grid, or the stock market), there are very few aspects of a normal day not mediated by computer systems. Two byproducts of this integration are increased frustration when these systems don’t work properly and corresponding appreciation for the value of dependable systems. The goal of my research is to build robust distributed systems that simply work despite the best, and worst, efforts of the environment. Building robust systems is challenging for multiple reasons. On the one hand, the definition of “works” is system dependent and frequently invokes a tension between performance and some notion of correct behavior. On the other hand, there are many environmental factors that can get in the way: crashed components, corrupted hardware, transient electromagnetic interference, delayed and corrupted messages, misconfigured servers, malicious attackers, disaffected sysadmins, selfish users, benign concurrency, changes in workload, . . . . My approach is based on a synergy between system building and theoretical work. To be confident that my systems work as advertised, I base my design and implementation efforts on solid theoretical foundations. To ensure that my theoretical foundations are relevant, I leverage the system building experience to develop new theoretical models and correctness conditions that better match the system deployment requirements. I have applied this methodology to a variety of applications in domains ranging from peer-topeer backup to live video streaming to scalable cluster file systems. In the process I have developed new fault models and built systems that are both more practical and more robust than previously thought possible. In the following I briefly overview of my previous work and future plans.1
منابع مشابه
Special Issue on Towards Quality Data with Fusion and Cleaning
Web users increasingly treat the Web as a very large knowledge base and use it to acquire all kinds of knowledge. In this article, we introduce a method that leverages the information on the Web to determine the truthfulness of any statement that attempts to state a fact. This method accomplishes the goal by checking whether there is an alternative statement that is more likely to be truthful. ...
متن کاملTruthfulness Analysis of Fact Statements Using the Web
Web users increasingly treat the Web as a very large knowledge base and use it to acquire all kinds of knowledge. In this article, we introduce a method that leverages the information on the Web to determine the truthfulness of any statement that attempts to state a fact. This method accomplishes the goal by checking whether there is an alternative statement that is more likely to be truthful. ...
متن کاملClement of Alexandria and His Doctrine of Scripture
This article discusses Clement of Alexandria’s doctrine of Scripture based on the terms “voice” of God, “first principle,” and “demonstration” that he uses for Scripture. By studying the usage of these terms and the related passages, it will be made clear that Clement regards Scripture as a vessel of God’s voice, favoring what is now called the theory of verbal inspiration. Mo...
متن کاملScience and ideology
More than 130 prominent Canadian physicians, scientists and public health professionals have endorsed this commentary. They are listed at: C ONSIDER THE FOLLOWING HYPOTHETICAL scenario. An innovative new intervention for people with diabetes is developed. Health Canada provides funding to a highly accomplished group of academic health scientists, who have no financial conflicts of interest with...
متن کاملNathan Clement December 30 , 2016 Research Statement Parabolic Higgs Bundles
My main project to date ([3]) is concerned with certain moduli spaces of Higgs bundles and isomorphisms between them, which arise from natural operations on the bundles. Conventionally1, a Higgs bundle is a pair (V, φ) on a Riemann surface X. Here V is a vector bundle (algebraic) and φ is a linear map V → ΩX⊗V called the Higgs field. Why consider such pairs? If nothing else, maybe a Higgs bundl...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 2013