tagging stream data for rich real-time...
TRANSCRIPT
![Page 1: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/1.jpg)
Tagging Stream Data for Rich Real-Time Services
Rimma V. NehmeElke A. Rundensteiner
Elisa Bertino
Presented by: Shujie Li13. 01. 2010
![Page 2: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/2.jpg)
Contents� Introduction
Approach Fundamentals
� Tag Model� Tag Query Language (TAG-QL)
Key StatementsAttach/Generate a tag to objects
� Tag-Base Query ProcessingTag-Oriented Query ProcessingTag-Aware Query Processing
� Experimental Analysis� Conclusion
![Page 3: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/3.jpg)
Contents� Introduction
ApproachFundamentals
��� Tag ModelTag ModelTag Model��� Tag Query Language (TAGTag Query Language (TAGTag Query Language (TAG ---QL)QL)QL)
Key StatementsKey StatementsKey StatementsAttach/Generate a tag to objectsAttach/Generate a tag to objectsAttach/Generate a tag to objects
��� TagTagTag---Base Query ProcessingBase Query ProcessingBase Query ProcessingTagTagTag---Oriented Query ProcessingOriented Query ProcessingOriented Query ProcessingTagTagTag---Aware Query ProcessingAware Query ProcessingAware Query Processing
��� Experimental AnalysisExperimental AnalysisExperimental Analysis��� ConclusionConclusionConclusion
![Page 4: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/4.jpg)
Introduction: Approach
Tag : Accident;2 Cars;
Near exit 12.Accident!!!!
Why stopped??
Accident or road
construction??
Query: Retrieve all
Tags within a
certain radius
![Page 5: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/5.jpg)
� Premise of tagging: Users can label data in order to get more informative query results.
� The additional label with type of metadata
� Continuous query processing with tags, address the Tick-tag issues and efficiency concerns.
Tick-tags
![Page 6: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/6.jpg)
Proposed Solution: Stream Tag Framework (STF)
TAG Query Language by Client side (User/Tagger) TAG-Based Query
Processing by Server side
Tick-tagTag results
Tuple results
Enriched tuple results
Data Model:
tuple = [ stream_id, tuple_id, A, timestamp ]
Data stream
![Page 7: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/7.jpg)
Introduction: Fundamental
1O
1O
Data Stream Management System (DSMS):The database for managing continuous data streams which are sequences of data tuples.
Tagging:The process of adding comments or labels to something.
Tagging in Data Stream Environments:Additional information to streaming objects (tuples, tuples attribute, etc.).
3o 2o 1o
1O
3o 2o 1o3o 2o 1o3t 2t 1t
streaming objects tagged stream
![Page 8: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/8.jpg)
Contents��� IntroductionIntroductionIntroduction
ApproachApproachApproachFundamentalsFundamentalsFundamentals
� Tag Model��� Tag Query Language (TAGTag Query Language (TAGTag Query Language (TAG ---QL)QL)QL)
Key StatementsKey StatementsKey StatementsAttach/Generate a tag to objectsAttach/Generate a tag to objectsAttach/Generate a tag to objects
��� TagTagTag---Base Query ProcessingBase Query ProcessingBase Query ProcessingTagTagTag---Oriented Query ProcessingOriented Query ProcessingOriented Query ProcessingTagTagTag---Aware Query ProcessingAware Query ProcessingAware Query Processing
��� Experimental AnalysisExperimental AnalysisExperimental Analysis��� ConclusionConclusionConclusion
![Page 9: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/9.jpg)
Tag Model: Fundamental
� Definition:Meta-Data tuples that attach additional information to stream objects.
� Characteristics:transient, sequential access, high input rate, pote ntially infinite size, continuous tag processing.
![Page 10: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/10.jpg)
Tag Model: Design
Tagger identifier
TID:
Unique id of the tagger / user, determined
by the system.
![Page 11: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/11.jpg)
Applicability
Applicability:
Describes the stream object, regular
expression.
![Page 12: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/12.jpg)
Content
Content:
A string datatype, stores the actual tag value.
E.g. “Accident”.
![Page 13: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/13.jpg)
Type
Type:
To classify streaming tags: Objective type
(i.e. “ 2 Car Accident” ), Subjective type (i.e. “Nice” , “ Interesting” ), etc…
![Page 14: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/14.jpg)
Sign
Sign:
To serve as a qualitative description of a
tick-tag based on the content to generate an
overall opinion for the tagged information.
“+” : Positive content; “ -” : Negative content.
![Page 15: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/15.jpg)
LifespanLifespan:
A time interval in which the tick-tag is
active.
Exception “ I” (Instant): if a single
applicability is wanted.
![Page 16: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/16.jpg)
ModeMode:
Indicates the user’s preference regarding the combination of the actual tag with
earlier ones.
“O” : Overwrite; “C” : Combine.
Point: Tagger specifies only self tags.
![Page 17: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/17.jpg)
Timestamp
Timestamp:
The time when the tick-tag was generated.
![Page 18: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/18.jpg)
Example:
Auction Stream contains items to sell
Schema: Seller_id Product Product feature
St_price time
Example:123 Dell
Laptoppink, 1420 600 Euro 2:00 pm
TAG1: with respect to VALUE of start price ( St_price ).
*,*, {St_price. value}
Fair + 1 day O
Value is given by system
![Page 19: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/19.jpg)
Contents��� IntroductionIntroductionIntroduction
ApproachApproachApproachFundamentalsFundamentalsFundamentals
��� Tag ModelTag ModelTag Model� Tag Query Language (TAG-QL)
Key StatementsAttach/Generate a tag to objects
��� TagTagTag---Base Query ProcessingBase Query ProcessingBase Query ProcessingTagTagTag---Oriented Query ProcessingOriented Query ProcessingOriented Query ProcessingTagTagTag---Aware Query ProcessingAware Query ProcessingAware Query Processing
��� Experimental AnalysisExperimental AnalysisExperimental Analysis��� ConclusionConclusionConclusion
![Page 20: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/20.jpg)
Tag Query Language:Key Statements
Syntax Meaning
ATTACH TAG... Attaches a tag to a streaming object
SELECT TAGS... Selects tags that satisfy a certain search predicate
SELECT TAGGED OBJECTS…
Selects tagged objects
SELECT…WITH TAGS
Returns tag-enriched query results
![Page 21: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/21.jpg)
How to attach a tick-tag to a streaming object ?
Method 1: manually attaching
Syntax:
ATTACH TAG <tag_content>
TO <object_description>
(WHERE <condition_description>)
(WITH
TAG_SIGH = <+ | - >
TAG_LIFESPAN = <lifespan_value>
TAG_MODE = <mode_value>)
Indicates the object to which the tag should be
attached
Decides the location of the tag
Tag Query Language:Attach a tag to objects
![Page 22: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/22.jpg)
Example1:
Tag with respect to VALUE of St_price
ATTACH TAG ‘ Fair ’
TO Auction . St_price . Value
WITH
TAG_SIGH = ‘+’ AND
TAG_LIFESPAN = 1 day AND
TAG_MODE =OVERWRITE
*,*, {St_price. value} Fair + 1 day O
![Page 23: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/23.jpg)
Keyword: Tagging is continuous
Continuous adding tick-tag
to the seller id’s value of auction with St_price > 600.
How to attach a tick-tag to a streaming object ?
Method 2: continuous attaching
Example2:
Tag with respect to VALUE of Seller_id
ATTACH TAG ‘ Expensive ’CONTINUOUSLYTO Auction . Seller_id . valueWHERE( SELECT Seller_id
FROM AuctionWHERE St_price > 600 )
WITHTAG_SIGH = ‘ - ’
![Page 24: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/24.jpg)
Contents��� IntroductionIntroductionIntroduction
ApproachApproachApproachFundamentalsFundamentalsFundamentals
��� Tag ModelTag ModelTag Model��� Tag Query Language (TAGTag Query Language (TAGTag Query Language (TAG ---QL)QL)QL)
Key StatementsKey StatementsKey StatementsAttach/Generate a tag to objectsAttach/Generate a tag to objectsAttach/Generate a tag to objects
� Tag-Base Query ProcessingTag-Oriented Query ProcessingTag-Aware Query Processing
��� Experimental AnalysisExperimental AnalysisExperimental Analysis��� ConclusionConclusionConclusion
![Page 25: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/25.jpg)
Category:
� Tag-Oriented Query Processing (TOQ Processing):
Users query tick-tags explicitly
� Tag-Aware Query Processing (TAQ Processing):
Users query tick-tags implicitly
Tag-Based Query Processing
![Page 26: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/26.jpg)
Expressing in TAG-QL:Q1: Tags where the tags values are of interest (‘select tags’)
Q2: Tags where the corresponding base data values are of interest (‘select
tagged object’)
pid measure location timePatient
Stream:
Q1: SELECT TAGS
FROM Patient
WHERE OBJECT =
Patient . measure AND
TAG_SIGN = ‘ - ’
Q2: SELECT TAGGED OBJECT
FROM Patient
WHERE TAG =‘ Emergency ’
Tag-Oriented Query Processing
![Page 27: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/27.jpg)
}]),,([
'T
o OtpOTO → with ttTt ii =∈∀ ,'
O2 satisfies po
Tag-Oriented Query Algebra:
Tagger Operator:
Input: a stream of objects & Output: a stream of objects
with an inserted tag t
Search predicate on objects
![Page 28: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/28.jpg)
Tag-Oriented Query Algebra:
Tag Selection:
Input: a stream objects with tags Output: a stream of tags
}]),([ 'TpOTS t
T
→ with TT ⊆'
t1 and t2 satisfy pt
Search predicate on tag
![Page 29: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/29.jpg)
}]),([ 'OpOTOS t
T
→
t1 and t2 satisfy pt
with OO ⊆'
Tag-Oriented Query Algebra:
Tagged Object Selection:
Input: a stream of objects with tags Output: stream of objects
![Page 30: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/30.jpg)
∅≠= ),( 21' TTET
} } }
]),,([
'21
'21
TTT
OEOOTJ →E: Some tag Join condition, i.e., if the both tags are equivalence, or if the both have the same meaning
t1 and t2 join based on E
with
Streaming
Tag Index
(stix)
42 11
Tag-Oriented Query Algebra:
Tag Join:
![Page 31: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/31.jpg)
} }
]),,([
'
'
T
G
aggT
T
aggT
OGEOTG →
t1 and t2 “are the same” based on E
Streaming Tag Index (stix)
42 11
State buffer Aggregated value
Tag-Oriented Query Algebra:
Tag-Based Aggregation:
Some tag join condition
A certain aggregate function
![Page 32: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/32.jpg)
Goals Tag-Aware Query:
Returns continuous query results that are “enriched” with
the tags attached to the original base data.
i.e. enriched tuples / tagged data tuples
Idea:
with statement “WITH TAGS”.
Tag-Aware Query Processing
![Page 33: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/33.jpg)
Expressing in TAG-QL:
pid measure location timePatient Stream:
Q3: SELECT pid, location, time
FROM Patient
WHERE measure > 80
WITH TAGS
![Page 34: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/34.jpg)
Projection operator:
Process tuples by extracting wanted attributes.
Propagates tick-tags and thereafter the projected tuples.
Discard the tick-tag which is attached to the projected attributes.
Tag-Aware Query Algebra:
![Page 35: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/35.jpg)
Example:
=Π )( 1, RcB
A B C
a b c
a b c
Stream R1
Relation Algebra Projection:
Tag-Aware Query Algebra:
B C
b c
tR1
{
Tuple
PO
SELECT B,C FROM R1
tb c
Data schema
}projected tuple
![Page 36: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/36.jpg)
Selection operator:
Drops tuples that do not satisfy the selection condition.
Propagation of tags delayed until min. one tagged tuple which
fulfills the selection condition found.
If all tagged tuples are filtered then their corresponding tag is
discarded.
Tag-Aware Query Algebra:
![Page 37: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/37.jpg)
Aggregation operator:
Each attribute domain is split into attribute sub-groups which
contain the tuples with the same attribute value.
A B C
a c d
b d e
b c f
tuple1
tuple2
tuple3
A B C
b d e
b c f
A B C
a c d
b c f
Sub-group from attribute A
Sub-group from attribute B
A B C
a c d
A B C
b d e
Tag-Aware Query Algebra:
![Page 38: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/38.jpg)
Aggregation operator:
Calculate a result for each sub-group.
Sends the result to the output stream preceded by the
collection of tags which are applicable to any object in that
sub-group.
Tag-Aware Query Algebra:
![Page 39: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/39.jpg)
Join operator:
Tuples x and y
join
(1: 1, 4-5)2 (1:1-3)1
Produces a new joined tuple
from x and y
Stream Tag Index
Tag-Aware Query Algebra:
![Page 40: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/40.jpg)
Contents��� IntroductionIntroductionIntroduction
ApproachApproachApproachFundamentalsFundamentalsFundamentals
��� Tag ModelTag ModelTag Model��� Tag Query Language (TAGTag Query Language (TAGTag Query Language (TAG ---QL)QL)QL)
Key StatementsKey StatementsKey StatementsAttach/Generate a tag to objectsAttach/Generate a tag to objectsAttach/Generate a tag to objects
��� TagTagTag---Base Query ProcessingBase Query ProcessingBase Query ProcessingTagTagTag---Oriented Query ProcessingOriented Query ProcessingOriented Query ProcessingTagTagTag---Aware Query ProcessingAware Query ProcessingAware Query Processing
� Experimental Analysis��� ConclusionConclusionConclusion
![Page 41: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/41.jpg)
Experimental AnalysisSetup:
� Stream Tag Framework is implemented in a DSMS prototype CAPE.
� Data generated by the Network-based Moving Objects Generator.
� 100K of moving objects, which present cars, cyclists, pedestrians.
� The moving objects stream are broken up into several streamsbased on the id of objects.
![Page 42: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/42.jpg)
Four Types of queries are used:
Tagger Operator
Tag Selection
Tag Join Tag-Aware Join
Query 1 Query 2 Query 3 Query 4
![Page 43: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/43.jpg)
Comparison:
Tagger Operator VS. regular Selection Operator
Tagger Operator is used!!
Tagger Operator is larger than regular Selection Operator !!!
![Page 44: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/44.jpg)
Comparison:
Tick-Tag VS Alternative Tagging Approach
Alternative Tagging Approach:
� Table ApproachProduce a separate global table which maintains all tags.
� Extended Data TuplesExtend the data tuple by adding an attribute for tag information.
� Streaming XML
Dynamic data which is in an XML format
![Page 45: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/45.jpg)
Comparison:
Tick-Tag VS. Alternative Tagging Approach
Tag Selection is used!!
Tick-Tag approach produces higher output rates !!!
![Page 46: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/46.jpg)
Tick-Tag approach produces smaller memory usage !!!
Tag Selection is used!!
Comparison:
Tick-Tag VS. Alternative Tagging Approach
![Page 47: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/47.jpg)
Cost of Tag Join Operator :
Tag Join is used!!
More tags, more overhead !!!!
![Page 48: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/48.jpg)
Comparison:
Tag-Aware Join VS. regular Join Operator
Tag-Aware Join is used!!
![Page 49: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/49.jpg)
Contents��� IntroductionIntroductionIntroduction
ApproachApproachApproachFundamentalsFundamentalsFundamentals
��� Tag ModelTag ModelTag Model��� Tag Query Language (TAGTag Query Language (TAGTag Query Language (TAG---QL)QL)QL)
Attach a tag to an object Attach a tag to an object Attach a tag to an object TagTagTag---Based Query ProcessingBased Query ProcessingBased Query Processing
TagTagTag---Oriented Query Processing (TOQP)Oriented Query Processing (TOQP)Oriented Query Processing (TOQP)TagTagTag---Aware Query Processing (TAQP)Aware Query Processing (TAQP)Aware Query Processing (TAQP)
��� Experimental AnalysisExperimental AnalysisExperimental Analysis� Conclusion
![Page 50: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/50.jpg)
Conclusion
� Propose the flexible STF to support for tagging data stream, and where the Tick-tags are attached to the objects.
� Tag Query Language enable attachment and query of streaming tags.
� Tag-Based Query Processing contains two aspects.
� Experiment shows the scalability and benefits of Tick-tags in contrast to the traditional theory.
![Page 51: Tagging Stream Data for Rich Real-Time Servicesresources.mpi-inf.mpg.de/departments/d5/teaching/... · Data Stream Management System (DSMS): The database for managing continuous data](https://reader035.vdocuments.net/reader035/viewer/2022070720/5ee08c27ad6a402d666bb693/html5/thumbnails/51.jpg)