Aurora

1
Aurora
Proponent Team Wei, Mingrui Liu,
Mo Rebuttal Team Joshua M Lee
Raghavan, Venkatesh
2
AURORA Proponent

• Introduction
• Motivation
• Aurora System Architecture
• Aurora System Query model
• Conclusion

3
IntroductionTraditional DBMSs

• Passive reporitory Human-Active,
  DBMS-Passive(HADP) model
• The current of state of the data is important
• Previous data needs to be extracted form the
  log
• Triggers and alerts as second-class citizens
• Perfect synchronization of data elements and
  exact query answers
• No real-time services from applications

4
Introduction Monitoring apps

• Monitoring applications are applications that
  monitor continuous streams of data
• Active repository DBMS-Active, Human-Passive
  (DAHP) model
• History of the data is important Not only the
  current state but also the previous history
• Triggers and alerts as the first-class citizens
• Missing or imprecise data, and approximate query
  answers
• Real-time services required by applications

5
Introducation Monitoring apps

• Target Applications military
• financial
  analysis
• tracking
• other
  real-time
• applications

6
Car Navigation System

• Data( e.g., the location of the car) comes from
  external sources
• History of the data is required( e.g., display a
  trajectory of your car in the past 20 minutes)
• Trigger and alert oriented an alert for the
  driver when the car is approaching to an
  intersection
• The location of the car is not always perfectly
  transmitted due to interferences etc..

7
Motivation-DSMS

• Data Stream Management Systems
• Streams continuous data feeds from sources.
  ect. sensors and satellites
• Monitoring applications track the data from
  numerous streams, filtering them for signs of
  abnormal activity and processing them for
  purposes of aggregation, reduction and
  correlation

8
Management Requirements
DBMS 1.Data processing results issuing
transactions and queries 2.Manages data in its
tables 3.Provide exact answers to exact queries
and is blind to real-time deadlines 4.Optimization
s of all queries in the same way 5.The norm is
pull-based queries
DSMS 1.Monitoring and alerting humans for
abnormal activities 2.Processing of data that is
bounded by finite window of values and not over
unbounded past 3.Respond to real-time deadlines
and provide reasonable approximations to
queries 4.Benefits from Application specific
optimization criteria (QoS) 5.The norm is
push-based data processing
9
What is Aurora

• Deals with large numbers of data streams
• Users built queries out of a small set of
  operators (boxes)
• Supports combination of (boxes) for better
  answers
• Aurora-Better support monitoring application
• Stream processing
• QoS functions
• Operators filtering, mapping, windowed
  aggregate, join
• Timeout

10
Aurora Architecture

• Continuous stream data comes
• Flow through a set of operators
• Output to application or materialized
• Multiple streams can be merged

11
Aurora System Query model
Supports continuous queries Supports
Ad-hoc queries
12
Aurora QoS Model

• ??Each output is supplied with a QoS
  specification
• ??QoS is captured by three functions
• ? ? A latency graph
• ?? A loss-tolerance graph
• ?? A value-based graph

13
QoS
14
Aurora optimization

• Dynamic continuous query optimization
• optimizer selects a portion of the network to
  optimize by
• Inserting projections.
• Combining boxes.
• Reordering boxes.

15
QoS data structures

• Response times output tuples should be produced
  in a timely fashion, as otherwise QoS will
  degrade as delays get longer.
• Tuple drops if tuples are dropped to shed load,
  then the QoS of the affected outputs will
  deteriorate.
• Values produced QoS clearly depends on whether
  or not important values are being produced.

16
QoS
17
Load Shedding

• Load shedding by dropping tuples
• Minimizing the overall performance degradation
  as a result of static analysis
• Semantic load shedding by filtering tuples
• Semantic load shedding based on value-based
  QoS information if available

18
Conclusion

• Aurora is a new rising star in DBMS
• More demand for monitoring applications
• Future directions
• Aurora for distributed processing
• More efficient data handing algorithm for
  missing and/or imprecise data that is common in
  sensor network

19
AURORA Rebuttal

• User Issues
• The graphical workflow style specification could
  be cumbersome in real life use
• An extended SQL may be easier for users