IMPLEMENTATIONS

1

• IMPLEMENTATIONS
• LOCUS file system OS
• TABS (Camelot) Data servers OS
• RAID Database level (server)
• SDD-1 Transaction manager
• Data manager
• System R Database level
• ARGUS Guardian (server)
• TABS section II, III C data server
• LOCUS set 3.4.3 - opening/reading file
• set 3.5 - synch. dist. file system activity
• ARGUS nested trans/guardians

2
CPU time used by RAID servers in executing
transactions (in seconds
3
(No Transcript)
4
Experiences with RAID Distributed Database

• Unix influences must be factored out.
• Communications software costs dominate everything
  else.
• Server based systems can provide modularity and
  efficiency.
• Concurrent execution in several server types is
  hard to achieve.
• Need very tuned system to conduct experiments.
• Data is not available from others for validation.
• Expensive research direction, but is respected
  and rewarded.

5
LAYERS

• Transaction
• Transparent to errors
• RPC executed within transaction
• Communication
• Datagram as basic unit
• RPC user level IPC
• (provides atomicity/reliability)
• Broadcast
• Input/Output System
• Lower level I/O
• (reading and writing raw data)
• Stable storage
• (log, UNDO, REDO)
• Upper level I/O
• (reliable read/write)

6
Description of Symbols
7
Times reported in seconds
8
The THOUSAND Relation
Some example tuples from the thousand relation
9
Execution times in 1-site RAID (in seconds)
10
Execution times in 2-site RAID (in seconds)
11
Execution times in 3-site RAID (in seconds)
12
Execution times in 4-site RAID (in seconds)
13
(No Transcript)
14
RAID Distributed System
DBMS
other applications
DBOS
other applications
RAID
OS
OS

• RAID supports reliability
• transactions
• stable storage
• buffer pool management

15
Figure 11. Performance comparison of the
communication libraries
16
RAID Transactions
17
Figure 2. Dependency graph for proposed layered,
distributed system
18
(No Transcript)
19
RESEARCH TASKS OVERVIEW
20
Time
(ms)
150
100
50
AD
AC
RC
CC
AM