2. Hbase Architecture
When the table becomes big then the table get
partitioned based on row key values into multiple
regions and regions are place where table data is
stored.
ID Name Location
11 Bob IN
22 Fara SG
33 Niki JP
44 Steve NZ
ID Name Location
11 Bob IN
22 Fara SG
33 Niki Jp
ID Name Location
44 Steve NZ
55 Nina RU
66 Ryan IN
Region 2
Region 1
Rupak Roy
3. Hbase Architecture
H Region Servers are responsible for splitting and
managing the regions and H Master is responsible
for managing and monitoring the cluster.
H Region Server
H Region
H Region
H Region Server
H Region
H Region
Rupak Roy
4. Hbase Architecture
Hbase uses Log Structured Merge, in short LSM tree to process to the
write operations.
In LSM all inputs are appended with a write head log which can be
reused to restore the database in case of any malfunctions.
Then the most recent inserts are
stored in the MemStorage and
when the memstorage gets filled
up the data is stored in a
new Hfile.
Hfile is a block indexed file format used by the Hbase to store the table
that is the key-value pairs on HDFS.
Hfile stores the key value in a lexical order by using thee row keys
MemStore
HFile
Write
Head
Log
H Region Server
Rupak Roy
5. Hbase Architecture
Some of the read and write operators are:
Get: gives the value of a Row Key
PUT: to insert a new entry.
Scan: gives the values for a range of Row Keys.
Delete: to delete a cell value.
Rupak Roy