get results, build your own big data beast : greenplum + dell
TRANSCRIPT
Get Results, Build Your Own Big Data Beast: Greenplum + Dell
Pivotal GreenplumDB
Master
Node Node Node Node Node Node
SCALE OUT NETWORKSCALE OUT NODES
MPP ( MASSIVE PARALLEL PROCESSING ) DB● Treat multiple physical databases as a single
logical database● Parallel databases utilize all the hardware
available to service queries● Standard SQL on massive data sets with
results in real time
R630 - Light and Fast
R730XD - Storage and IO Master
R830 - Processing Powerhouse
Dell servers create a monstrous platform of capabilities, clusters can be tuned for specific use cases.
Simple Architecture
Master
Nodes
On Standard Enterprise Hardware
Parallel Resource Utilization
Master
Nodes
SELECT * FROM MassiveTable
SELECT * FROM MassiveTable
SELECT * FROM MassiveTable
SELECT * FROM MassiveTable
SELECT * FROM MassiveTable
SELECT * FROM MassiveTable
SELECT * FROM MassiveTable
SELECT * FROM MassiveTable
SELECT * FROM MassiveTable
SELECT * FROM MassiveTable
SELECT * FROM MassiveTable
SELECT * FROM MassiveTable
SELECT * FROM MassiveTable
SELECT * FROM MassiveTable
SELECT * FROM MassiveTable
SELECT * FROM MassiveTable
SELECT * FROM MassiveTable
Leverage The Full Power Of The Hardware Stack
Scales As Hardware Is Added
Master Nodes
Expand To Meet Resource Needs
Not Just A Database
A Data Science Toolkit
Use powerful languages on data in parallel.
Machine Learning implemented in SQL
Driving Results for Big Data Leaders For Years On Software Recently Open Sourced
Baseline Sample Architecture*
Interconnect -Dual 10G Bonded
S4048-ON
MasterStandby Master
Node1
Node2
Node3
Node4
Node5
Node6
Node7
Node8
R730xd2xE5-2650v424x1.8TB256GB RAMH730P
~100TB of DB data~400TB w/ compression
To ExternalNetwork
*Architecture easily modified to fit needs