Batch vs Continuous Graphs in Abinitio - A Revision
Batch vs Continuous Graphs in Abinitio - A Revision
watch my video on Batch versus Continuous Graphs
Batch vs Continuous Graphs in Abinitio - A Revision
We will talk about below-
a. Batch versus Continuous
long running , wait | continuous running , real time response
best for analytical application | best for real time transactional applications
run limited no of times a days | can run forever of can shutdown/stop at specified time
b. Type Of Continuous Graphs
a. request/response
b. publish/subscribe
c. Continuous Graph is used when?1. input is not a fixed set of data but a regular flow , perhaps coming at rates near real time
2. results are need to be made available before the large batch of data accumulate; low latencies
are desired
3. the overhead of graph startup and shutdown to process small batch of incoming data dominates
processing time
4. recovery model suggests it - you need to recover based on group of records,
not based on groups of components ( phases) this can be case the amount of work per record is
huge.
5. your application design is inherently continuous , for example service graph
for example,
SORT , sorts the records within each unit of work.
Roll Up (in-memory), produces the output based on each unit of work
Data base components , data base commit happens at the end of each unit of work
For More Abinitio, AWS, Database content please visit my youTube channel.
https://www.youtube.com/@datapundit
@datapundit
Comments
Post a Comment