Impala and BigQuery (1)

8/12/2019 Impala and BigQuery (1)

http://slidepdf.com/reader/full/impala-and-bigquery-1 1/47

Impala and BigQuery

By David Gruzman

BigDataCraft.com



Impala and BigQuery

by David Gruzman

► Big Query is google's database service basedon the Dremel. Big Query is hosted by Google.

►Impala is open source database inspired by the

Dremel paper. Impala is part of the ClouderaHadoop distribution.



Today agenda

► Overview of Dremel as a technology

► Overview of the BigQuery

► A few words about Impala

► DG Mediamind use case

► Deeper insights into Impala

► Conclusions► Q&A



Why dremel?

► Google is first who got MapReduce

► Google is first faced MapReduce main problem – latency. The problem was propagated to

engines on top of MapReduce also.► It is logical that Google was first who

approached it by developing real time query

capability for big data.



How dremel is used in google

► Dremel is not replacement for the MapReduceor Tenzing but complements it. (Tenzing isGoogle's Hive)

► Analyst can make many fast queries usingDremel

► After getting good idea what is needed – runslow MapReduce (or SQL based onMapReduce) to get precise results



Why dremel is Unique

► Dremel with BigQuery built on top of it isprobably only Interactive big data query enginetoday.

► I mean that it is only engine capable to produceresults over terabytes of data in seconds!

► Main idea (my guess) that is harness huge

cluster of machines for the single query.



Dremel as technology

Novel Hierarchical columnar format.

► LLVM based code generation.

► Distributed aggregation Tree

► In-situ data processing. (inside the storage)



Dremel : Aggregation tree



Dremel : Nested columnar format



Big Query

► Service built by google on top of the Dremelengine

► Only (known to me) query engine as a service

working with BigData.► Query time not depends on data size



BigQuery main capabilities

► Aggregations

► Join of big table to small table.

► Join of two big tables (recently added)

► Hierarchical data format. It makes pre-aggregations cheaper.



Main limitations

► Small results size

► Intermediate results should not exceed memorysize.

► No “external tables”



Why BigQuery is not popular



So,why BigQuery is not popular

► Data is not created in google cloud. It is hardand not practical to move big data. It is heavy,after all.

► Google is used to change APIs. BigQuery alsochanged during last years. It is hard to buildbusines.

► Many companies in Internet related businessesa wary of sharing data with Google.

► It is expensive. 35$ per TB can give 1000th ofdollars bills per day.



Dremel



In the same time – it is goodtechnically

► I got referances from company doing serioustesting

► Marting Fawler's company also tested it and

give very good feedback.



Question to all of you

Why Your organization decided not to usegoogle's Big Query?



Where we can find Impala





What is impala

► Massive parralel processing (MPP) databaseengine, developed by Cloudera.

► Integrated into Hadoop stack on the same level

as MapReduce, and not above it (as Hive andPig)

HDFS

Map Reduce

HivePig

Impala



Why impala

► Data has a gravity

► Today a lot of data live in HDFS

► It is not practical to move big data

► It is practical to bring engine to the data

► In the same time – MapReduce is not must

►

Impala process data in Hadoop cluster without using MapReduce



MapReduce bypass

► Several other modern Database engines alsorealized the opportunity to bypass MapReducebut work right with HDFS.

► They takes various approaches.►



MapReduce Bypass

► Existing MPP databases, like Greenplum – store their external tables in the HDFS



MapReduce bypass

► Jethrodata store data in their own format onHDFS and also work with it without MR layer.

► They have their proprietary format which enable

full indexing of the data together with columnarefficiency. In cases of high selectivity queriesthis approach has serious advantages.



Use Case from DG

I think it is will be typical case in the future

► DG is using Hadoop and Hive

► Evaluation Impala to do part of things more

efficiently.

► After their case presentation we will back todiscuss insights of the Impala

A i I l h diff t l



Again – Impala has different placethen Pig and Hive

HDFS

Map Reduce

Hive and Pig

Impala



Impala architecture



Impala – Dremel traces

► LLVM code generation

► It is really fast

► C++ as implementation language (not Java...)

► Simple query engine. It actually doing thingswhich can be done in memory.

► Broadcast join algorithm is implemented



LLVM code generation

► Assume you want to write custom code for thespecific query. It will be super efficient

► Code generation automate this process for

each query► We actually need to super-optimize inner loop

doing filtering (where) and group by.

► LLVM enables us to compile in fraction ofseconds into native code

► LLVM enable us to enjoy new CPU capabilitieslike SSE in a portable way.



Why code generation it interesting?

► If you develop own engine, or some peace ofcode responsible to process serious datavolumes code generation may give you order ofmagnitude boost.

► I had cases when usage of such technologywas game changing



Impala – Hive Traces

► While dremel converts data into own format,Impala supports multiple formats. It is kind ofschema on read.

► Impala shares metastore with Hive, whichenables very simple adoption

► Internally Impala have well defined way to addnew formats





Impala vs MPP

► It usually tooks many years to create MPPdatabase.

► There are serious simplifications:

► The data is read only

► There is actually not DBMS – only queryengine.

► No serious resource management, butmeasurement (all over code).



Impala – hive killer?

► Not so quickly.

► Hive is doing things Impala can not do yet, like joins between several big tables.

► Hive has convinient java UDF, while impala isnot

► Impala does not have inter-query fault

tolerance.► In the same time – MapReduce is not good

framework for the database engine



Impala – Data Formats

► There are scanners for the following types:

► RCFile

► Parquet (native dremel format)

► CSV

► AVRO

► Sequence File



Impala – future

► Will get closer to other MPP engines

► Support more formats

► More advanced scheduling and resource

management



Basic benchmark

► TPC-H, Q1, SF=10

► 4 EC2 large instances

► 4 seconds, while hive takes about 1 minute.

► This number means group by speed of about235MB/sec per core.



Impala price per GB

► 1 Large instance costs $0.24

► Cluster costs 0.96 per hour.

► Cost of 1 second : 0.96 / 3600

► We process by such cluster 1.75GB per second

► So cost of 1 TB processing is about $0.15

► It is about 300 times cheaper then BigQuery





What with clouds?

I l i l d i t l ti



Impala in cloud is not elastic

► To be elastic we need to create cluster whenwe need it.

► Even if we agree to by hour resolution – storage

will be a problem► S3 will not give us hundreds of Mbs per second

per instance

►

To store data in local file system – is transient

I l l i



Impala - conclusions

► It is first time I remember when we can put ourhands on free MPP database.

► There is no risk to try it side-by-side with Hive

► It is possible to offload part of the work toImpala and do the rest with Hive

► It is part of the Cloudera Hadoop distribution

and easily installed by Cloudera Manager

M t i l d



Materials used

► Benchmarks

http://www.slideshare.net/sudabon/performance-evaluation-of-cloudera-impala-20121208-

15536323https://amplab.cs.berkeley.edu/benchmark/

► Architecture

http://www.slideshare.net/scottleber/impala-19176906

https://cloud.google.com/files/BigQueryTechnicalWP.pdf

M t i l d i

http://www.slideshare.net/sudabon/performance-evaluation-of-cloudera-impala-20121208-15536323

























Material used - comparisons

► To hive: http://www.quora.com/Cloudera/Does-Cloudera-Impala-have-any-drawbacks-when-compared-with-Hive

► To vertica: http://www.quora.com/Cloudera-Impala/How-does-Cloudera-Impala-compare-to-Vertica

► To dremel: http://www.quora.com/Cloudera-

Impala/How-does-Clouderas-Impala-compare-to-Googles-Dremel

Th k !!!

http://www.quora.com/Cloudera-Impala/How-does-Cloudera-Impala-compare-to-Vertica





















Thank you!!!

► Special thanks to

► Faina Kamenetsky – who helped set up clustersin amazon.

BigDataCraft com



BigDataCraft.com

► We are boutique consulting company► Our services are:

► On paper POC

► On hardware POC

► Architecture / Design reviews

► Custom integrations and bug fixing

Impala Flow



Impala - Flow

Impala and BigQuery (1)

Documents