N.Tabul, U.Cetintemel, S.Zdonik, M.Cherniack, M.Stonebraker "Load Shedding in a Data Stream Manager", VLDB 2002
Problem:
Load Shedding algorithm. Experiment implemented using RR CPU scheduling policy. Monitoring process which wakes up periodically to check the load.
Questions:
1. What type of operators are we talking about? Does it matter ?
2. Never discuss about how the cost of adding boxes can be estimated
3. Join operator only works for long window sizes. For small ones the values cannot be assumed to be uniform and selectivity estimation becomes a problem.
Labels
annotation
(10)
web service
(3)
uncertainty
(2)
API
(1)
QoD
(1)
bio
(1)
confidence intervals
(1)
data streams
(1)
grammar
(1)
load shedding
(1)
load shedding ; aggregate queries
(1)
load shedding ; continuous queries ;
(1)
load shedding ; dynamic data streams
(1)
meta-scheduling
(1)
monitor dropped packets
(1)
multi-values
(1)
online
(1)
propagation
(1)
provenanace
(1)
punctuation; security
(1)
record matching
(1)
review_paper
(1)
semantics;
(1)
staleness
(1)
Tuesday, September 23, 2008
data streams; survey paper
Brian Babcock, S. Babu, M. Datar, R. Motwani, J. Windom "Models and Issues in Data Stream Systems", ACM 2002
Friday, September 12, 2008
load shedding for aggregation queries, ICDE2004
B. Babcock, M. Datar, R. Motwani, "Load Shedding for aggregation Queries over Data Streams", ICDE 2004
Problem:
Which tuples to drop, and where in the query plan to drop them, so that the degree of inaccuracy in the query answers introduced as a result of load shedding is minimized.
Model: a query made by multiple operators that ends with an aggregate operator.
Solution:
Seen as an optimization problem: minimizing inaccuracy in query answers, subject to the constraint that system throughput must match or exceed the data input rate.
Uses Statistics Manager to keep track of mean of variance of the values in the input tuples.
QoS: add weights on the queries and use it to compute the relative error introduced by the load shedder.
Questions:
1. How is the load shedding done?
Problem:
Which tuples to drop, and where in the query plan to drop them, so that the degree of inaccuracy in the query answers introduced as a result of load shedding is minimized.
Model: a query made by multiple operators that ends with an aggregate operator.
Solution:
Seen as an optimization problem: minimizing inaccuracy in query answers, subject to the constraint that system throughput must match or exceed the data input rate.
Uses Statistics Manager to keep track of mean of variance of the values in the input tuples.
QoS: add weights on the queries and use it to compute the relative error introduced by the load shedder.
Questions:
1. How is the load shedding done?
Tuesday, September 9, 2008
security in data streams
R. Nehme, E. Rundensteinerr, E. Bertino, "A security Punctuation Framework for Enforcing Access Control on Streaming Data", IEEE International Conference on Data Engineering, (ICDE), Cancun, Mexico, April 2008
Problem:
Privacy and security in the context of data stream management systems (DSMS)
Solution:
Security restrictions are streamed together with data. Extend CQL to support specifications of security punctuations.
Problem:
Privacy and security in the context of data stream management systems (DSMS)
Solution:
Security restrictions are streamed together with data. Extend CQL to support specifications of security punctuations.
Wednesday, September 3, 2008
data streams: semantics
D.Maier, J. Li, P. Tucker, K. Tufte, V. Papadimos, "Semantics of Data Streams and Operators", ICDT 2005, Springer-Verlag 2005
Problem:
1. interpretation of a stream might vary from application to application. Need a mechanism for expressing and reasoning about a stream semantic and representation.
2. needs ways to express the information about content and presentation order.
Solution:
1. Reconstruction Function; maps each prefix, P, of a stream into the desired domain of interpretation
2. stream is viewed as a mixture of finite sub-streams; the ends of a sub-stream can be determined
Questions left:
2. how to determine the end of a sub-stream? What if it's not possible?
Problem:
1. interpretation of a stream might vary from application to application. Need a mechanism for expressing and reasoning about a stream semantic and representation.
2. needs ways to express the information about content and presentation order.
Solution:
1. Reconstruction Function; maps each prefix, P, of a stream into the desired domain of interpretation
2. stream is viewed as a mixture of finite sub-streams; the ends of a sub-stream can be determined
Questions left:
2. how to determine the end of a sub-stream? What if it's not possible?
Subscribe to:
Posts (Atom)