Presented by: Mingzhu Wei and Abhishek Mukherji

1
Presented by Mingzhu Wei and Abhishek Mukherji
2
Motivation
3
Categories of Location-aware Queries

• Query Stationary
• Object Moving

4
Snapshot vs. Continuous Query Processing

• Traditional Spatio-Temporal (Snapshot) Queries

Data
5
Challenge I Massive size of incoming data streams

• Spatio-temporal Databases
• secondary storage
• DSMS
• Load Shedding
• Immediately drop insignificant tuples
• Possibly reduce the query area
• Objects that satisfy less than k queries are
  insignificant
• Lazily drop insignificant tuples
• PLACE
• Predicate-based window

6
Challenge II Continuous evaluation of CQ

• Spatio-temporal Databases
• Associate a validation condition with each query
  answer

• Valid time (t)
• The query answer is valid for the next t time
  units
• Valid region (R)
• The query answer is valid as long as you are
  within a region R

• It is challenging to maintain the computation of
  valid time/region for querying moving objects
• DSMS
• Sliding-window
• PLACE
• Progressive evaluation paradigm

7
Challenge III Wide variety of query types

• Spatio-temporal Databases
• Have solution for stationary range queries on
  moving objects such as aggregation and k-NN
  queries
• DSMS
• Wide range of query operators
• No operators for spatio-temporal queries
• PLACE
• Extend PREDATOR and NILE SQL language
• Use of INSIDE and kNN operators

8
Challenge IV Large number of concurrent queries

• Spatio-temporal Databases
• Solution
• Centralized environment Q-index (R-tree-like)
• Distributed environment Ship part of query
  processing down to the moving objects to save
  server from bottleneck
• DSMS
• Multi-query Optimization techniques
• Sharing query plan
• Sharing at operator level
• PLACE
• Shared execution paradigm Spatial Join

9
PLACE Architecture
10
Data Models

• Three-level Storage Hierarchy
• In-memory
• Subset of incoming data stored in memory
• Associate with outstanding queries
• Cache readings
• Cache readings and flush them to secondary
  storage
• Secondary storage
• Sample data and choose kth reading to disk
• Keep one reading of objects and queries
• Index data using grid structure
• Repository storage
• Take snapshot of in-disk database every Tarchive
  time
• Multi-version structure of moving object

11
Extended SQL Syntax

• inside_clause
• Stationary query (x1,y1,x2,y2)
• Moving query (M,OID, width, length)
• knn_clause
• Stationary query (k,x,y)
• Moving query (M, OID, k)

Q1 What is the query and the object here? Q2
Query result for K2?
12
More Questions!

• The first k objects are considered an initial
  answer
• Q3 Is there any relation between the k-NN query
  and the range query?
• K-NN query is reduced to a circular range query

However, the query area may shrink or grow
K 3
13
Predicatebased Sliding Windows

• Temporal expiration
• Same as sliding window
• Spatial expiration
• Predicate-based expiration
• Other form of predicates

14
Predicate-based Sliding Window (continued)

• Only significant objects are stored in-memory
• An object is considered significant if it is
  either in the query area or the cache area

• Due to the query and object movements, a stored
  object may become insignificant at any time
• Larger cache area indicates more storage overhead
  and more accurate answer

15
Predicate-based Sliding Window (continued)Cachin
g the Result

• Observation Consecutive evaluations of a
  continuous query yield very similar results
• Idea Upon evaluation of a continuous query,
  retrieve more data that can be used later

• K-NN query
• Initially, retrieve more than k
• Range query
• Evaluate the query with a larger range

• How much we need to pre-compute?
• How do we do re-caching?

16
Incremental Evaluation

• The query is evaluated only once. Then, only the
  updates of the query answer are evaluated
• There are two types of updates- Positive and
  Negative updates

Query Result

• Only the objects that cross the query boundary
  are taken into account

• Need to continuously listen for notifications
  that someone cross the query boundary

17
Spatio-temporal Incremental Pipelined Operators

• Pipelined Query Operators
• Combination of Spatio-temporal operators with
  regular CQ operators
• ST Operators Pushdown- reduces the number of
  tuples
• Flexible Query Optimization- multiple candidate
  execution plans
• Algorithm
• Keep track of recently reported answer Q.Answer
  of each query Q
• For each new coming tuple P, test
• Is P part of previously reported Q.Answer?
• Does P qualify to be part of the current answer?
• Four cases
• CASE I P is part of Q.Answer and P still
  qualifies- P will not be processed
• CASE II P is part of Q.Answer and P does not
  qualifies- negative P propagated
• CASE III P is not part of Q.Answer and P
  qualifies- positive P propagated
• CASE IV P is not part of Q.Answer and P does not
  qualifies- P has no effect on Q

18
Scalability

• Continuous queries last for long times at the
  server side
• While a query is active in the server, other
  queries will be submitted
• Shared execution among multiple queries
• Should we index data OR queries?
• Data and queries may be stationary or moving
• Data and queries are of large size
• Data and queries arrive to the system with very
  high rates
• Treat data and queries similarly
• Queries are coming to data OR data are coming to
  queries?
• Both data and queries are subjected to each other
• Join data with queries

19
ScalabilitySpatial Join

• To accommodate for the continuous movement of
  both data and queries
• Concurrent continuous queries share a grid
  structure
• Moving objects are hashed to the same grid
  structure as queries
• The spatio-temporal join is done by overlaying
  the two grid structures

20
Scalability

• Evaluating a large number of concurrent
  continuous spatio-temporal queries is abstracted
  as a spatio-temporal join between moving objects
  and moving queries

21
Performance Evaluation

• Size of incremental answer
• Pipelined spatio-temporal operators
• Pipeline with a select operator
• Pipeline with a join operator

22
Conclusion

• PLACE Pervasive Location-Aware Computing
  Environments
• Scalable execution of continuous queries over
  spatio-temporal data streams
• Shared execution among concurrent continuous
  queries

• Built inside a database engine
• Incremental evaluation of continuous queries
• Spatio-temporal query operators
• Treats Query and Data Symmetrically
• Out-of-order expiry of tuples