RAD Lab: PIQL / Modeling Query Response Time
246 views · Published 11 March 2011 · 21:35 · Indexed 23 September 2026
Channel: CITRIS and the Banatao Institute · 2011 · Science & Technology
Speaker: Kristal Curtis Abstract: PIQL, the Performance Insightful Query Language, is a library that allows developers to write declarative queries that can be broken down into operations against a key-value store. In a climate where many developers are abandoning traditional RDBMSs in favor of NoSQL stores for their predictable performance, PIQL aims to retain some of the advantages of SQL -- namely, a declarative query language, physical/logical data independence, and optimizer-generated query plans -- while still providing performance safety by requiring that all queries be annotated by a fixed bound on the amount of data that they will process. Since this bound must be independent of the amount of data in the system, this guarantees that the queries will not become unmanageably slow when the user base grows. Merely setting a bound, however, is not enough; an unreasonably large value will still result in unsatisfactory performance as the site scales. Therefore, PIQL also provides a Performance Insight Assistant that guides developers to choose bounds that are consistent with their site's service level objective (SLO).
More from this channel
-
42:43
Participatory Urbanism: Empowering Everyday Civic Engagement
-
1:09:44
From Zero to Fifty in Two Years: An Update on Microsoft Rese
-
51:04
Using Mathematical Models to Help Understand Planar Cell Pol
-
54:42
Cap and Trade Programs, and Climate-Friendly Innovation
-
59:11
Engineering Nano/Biological Interfaces
-
46:19
Ambulatory Sculptures and Dancing Robots
-
52:13
Policy and Economic Issues in Dynamic Electricity Pricing
-
57:58
Towards Personal, Proactive, & Preventive Healthcare: Disrup