before @werner’s keynote, I’m going to sneak in, “Deep dive on Amazon Neptune” (DAT309) by @b2ebs

because otherwise, I’m going to make @joeyskladany

#reinvent

“Graphs are awesome”, @b2ebs << nails it right out of the gate

#reinvent

despite being a 300-level talk, graph 101 is entirely appropriate here

a lot of ppl need this introduction to realize what this tool is good at

#reinvent

image is air traffic routes…at least old air traffic routes

#reinvent

some cases where graphs are a good choice:

- looking at patterns in relationships
- traversing relationships

#reinvent

before you all jump in & say “I can do this in an RDBMS or ___” …yes, yes you can

it’s a question of the experience and performance of the analysis. building a question to explore these types of datasets in standard SQL is a nightmare…but doable

simple via graphs

#reinvent

identity graphs are another solid use case

#reinvent

ditto for fraud

#reinvent

lots of customers using Amazon Neptune for graphs today << probably not as many as it would help though. lots of work to do around graph education

#reinvent

here are some of the key tools and projects you’ll use with graphs

#reinvent

Amazon Neptune has a lot going on under the covers

#reinvent

the notebook piece is huge. a great way to get introduced to graph concepts and to make sure that you’re accomplishing your goals with Neptune

check out the @AWSOpen project for those notebooks at https://github.com/aws/graph-notebook

#reinvent

the streams feature of Neptune is _extremely_ useful. learn more at https://docs.aws.amazon.com/neptune/latest/userguide/streams.html

#reinvent

. @awscloud Neptune also provides the ability to ask what your queries are doing. learn more about that at https://docs.aws.amazon.com/neptune/latest/userguide/streams.html

#reinvent

underneath the covers of Amazon Neptune; all of the @awscloud reliability and scalability features you’d expect

#reinvent

more on the plumbing for Neptune

#reinvent

lots of technical details in this session on @awscloud Neptune, exactly what we want in a 300-level talk

#reinvent

back to the air routes dataset for some example walkthroughs...

#reinvent

notice the simplicity in the logic of the query? that’s a huge advantage of graphs << when picking a data store, you really need to think about how you use it. do you write more? simple reads? complex queries? that’ll point you in the right direction

#reinvent

here’s what that data looks like on disc in the data model

it’s simple to write to the store but the graph db engine allows for that complicated query to be written simply and run very fast

#reinvent

here’s the SparQL query for the same question. a little bit clunkier but still effective

#reinvent

more on SPARQL: https://jena.apache.org/tutorials/sparql.html

more on Gremlin: https://tinkerpop.apache.org/gremlin.html 

#reinvent

SPARQL data model in @awscloud Neptune. not that different from the TinkerPop one

#reinvent

…honestly that’s probably a huge challenge for most ppl getting into graphs. what’s the difference between TinkerPop, Gremlin, SPARQL? why do I care? what are the non-reversible choices I have to make right now?

#reinvent

. @awscloud Neptune does a good job of limiting the impact of these choices and hiding the complexity BUT the larger graph community has a lot of work to do to in this area

#reinvent

nice tip here. you can use AWS Database Migration Service to load data into Neptune

learn more at https://docs.aws.amazon.com/neptune/latest/userguide/dms-neptune.html

#reinvent

related sessions for Neptune at #reinvent

solid talk from @b2ebs on @awscloud Neptune. learn more about the service at https://aws.amazon.com/neptune/ 

#reinvent