Global Positioning System Reference
In-Depth Information
Related Work
We fi rst introduce the data model and the basic operations of HBase and
Cassandra. We then present the traditional multi-dimensional indexing
techniques, linearization and index trees. In addition, we illustrate the
existing multi-dimensional indexing techniques, RT-CAN and MD-HBase,
developed for CDMs.
HBase and Cassandra
Data Model
HBase and Cassandra adopt the BigTable-like data model, which is a
column-oriented data model. The data model of BigTable is constituted
of columns, where each column is expressed as (name, value, timestamp),
where a timestamp is an updated time of a column. For HBase, the data is
structured in tables, row keys, column families, and columns. Specifi cally,
each table comprises row keys and column families; and each column family
contains one or more columns; each row consists of a key and columns
mapped by the key. Given the data in Table 7, HBase stores the data as
Table 6. For example, for key p 1 , the corresponding data include three
column families, and the second column family comprises two columns
whose names are rest.name and rest.lng. In addition, for each key in HBase,
the corresponding data can own multiple versions, identifi ed by different
timestamps. For instance, restaurant p 1 was checked by user u 1 and user u 2
on 8 May 2011 and on 8 August 2011, respectively. In addition, in HBase,
the keys and columns' names are stored in lexicographical order.
Compared with HBase, the data stored in Cassandra are structured in
two ways, column family or super column family. In Cassandra, a column
family consists of keys and columns mapped by the keys, and a super
column family consists of keys and the corresponding super columns, in
which each super column is expressed by (super column name, columns).
For instance, given Table 7, the data in Cassandra can be stored as Table 8a
and b, which are column families. In addition, an example of a super column
Table 6. A table of HBase.
timestamp
column
family:
column family:
column family:
keys
rest.name
rest.lan
rest.lng
uid
2011/05/08
Friday
24.805
120.995
u 1
p 1
2011/08/08
Friday
24.805
120.995
u 2
2011/08/30 McDonald's
24.794
121.002
u 2
p 2
2011/10/10 McDonald's
24.794
121.002
u 3
p 3
2011/11/07
KFC
24.794
121.005
u 4
Search WWH ::




Custom Search