Towards Parameter-Free Data Mining - Keogh et al. SIGKDD 2004 Another time series paper today from the Facebook Gorilla references. Keogh et al. describe an incredibly simple and easy to implement scheme that does surprisingly well with clustering, anomaly detection, and classification tasks over time series data. As per the title of the paper, it … Continue reading Towards parameter-free data mining
Year: 2016
Finding surprising patterns in a time series database in linear time and space
Finding Surprising Patterns in a Time Series Database in Linear Time and Space - Keogh et al. SIGKDD 2002 In the Facebook Gorilla paper, the authors mentioned a number of additional time series analysis techniques they'd like to add to the system over time. Today's paper is one of them, and it deals with the … Continue reading Finding surprising patterns in a time series database in linear time and space
NOVA: A Log-Structured File System for Hybrid Volatile/Non-Volatile Main Memories
NOVA: A Log-structured file system for hybrid volatile/non-volatile main memories - Xu & Swanson 2016 Another paper looking at the design implications of mixed DRAM and NVMM systems (it's the future!), this time in the context of file systems. (NVMM = Non-volatile Main Memory). Hybrid DRAM/NVMM storage systems present a host of opportunities and challenges … Continue reading NOVA: A Log-Structured File System for Hybrid Volatile/Non-Volatile Main Memories
Uncovering bugs in Distributed Storage Systems during Testing (not in production!)
Uncovering bugs in Distributed Storage Systems during Testing (not in production!) - Deligiannis et al. 2016 We interviewed technical leaders and senior managers in Microsoft Azure regarding the top problems in distributed system development. The consensus was that one of the most critical problems today is how to improve testing coverage so that bugs can … Continue reading Uncovering bugs in Distributed Storage Systems during Testing (not in production!)
BTrDB: Optimizing Storage System Design for Timeseries Processing
BTrDB: Optimizing Storage System Design for Timeseries Processing - Anderson & Culler 2016 It turns out you can accomplish quite a lot with 4,709 lines of Go code! How about a full time-series database implementation, robust enough to be run in production for a year where it stored 2.1 trillion data points, and supporting 119M … Continue reading BTrDB: Optimizing Storage System Design for Timeseries Processing
Gorilla: A fast, scalable, in-memory time series database
Gorilla: A fast, scalable, in-memory time series database - Pelkonen et al. 2015 Error rates across one of Facebook's sites were spiking. The problem had first shown up through an automated alert triggered by an in-memory time-series database called Gorilla a few minutes after the problem started. One set of engineers mitigated the immediate issue. … Continue reading Gorilla: A fast, scalable, in-memory time series database
Slacker: Fast Distribution with Lazy Docker Containers
Slacker: Fast Distribution with Lazy Docker Containers - Harter et al. 2016 On you marks, get set, docker run -it ubuntu bash. How long did it take before you saw the bash prompt? In this wonderful FAST'16 paper, Harter et al. analyse what happens behind the scenes when you docker run a container image, and … Continue reading Slacker: Fast Distribution with Lazy Docker Containers
Optimizing Distributed Actor Systems for Dynamic Interactive Services
Optimizing Distributed Actor Systems for Dynamic Interactive Services - Newell et al. 2016 I'm sure many of you have heard of the Orleans distributed actor system, that was used to build some of the systems supporting Microsoft's online Halo game. Halo Presence is an interactive application which implements presence services for a multi-player game running … Continue reading Optimizing Distributed Actor Systems for Dynamic Interactive Services
Data Tiering in Heterogeneous Memory Systems
Data Tiering in Heterogeneous Memory Systems - Dulloor et al. 2016 Another fantastic EuroSys 2016 paper for today, and one with results that are of great importance in understanding the cost and performance implications of the new generation of non-volatile memory (NVM) heading to our data centers soon. Furthermore, we also get some great insight … Continue reading Data Tiering in Heterogeneous Memory Systems
GeePS: Scalable deep learning on distributed GPUs with a GPU-specialized parameter server
GeePS: Scalable deep learning on distributed GPUs with a GPU-specialized parameter server - Cui et al. 2016 (EuroSys 2016) We know that deep learning is well suited to GPUs since it has inherent parallelism. But so far this has mostly been limited to either a single GPU (e.g. using Caffe) or to specially built distributed … Continue reading GeePS: Scalable deep learning on distributed GPUs with a GPU-specialized parameter server
the morning paper 