HBaseCon, profile picture
Uploaded byHBaseCon
PPTX, PDF20,062 views

A Survey of HBase Application Archetypes

Lars George and Jon Hsieh presented archetypes for common Apache HBase application patterns. They defined archetypes as common architecture patterns extracted from multiple use cases to be repeatable. The presentation covered "good" archetypes that are well-suited to HBase's capabilities, such as storing simple entities, messaging data, and metrics. "Bad" archetypes that are not optimal fits for HBase included using it as a large blob store, naively porting a relational database schema, and as an analytic archive requiring frequent full scans. A discussion of access patterns and tradeoffs concluded the overview of HBase application archetypes.

Embed presentation

Downloaded 478 times
Headline Goes Here Speaker Name or Subhead Goes Here DO NOT USE PUBLICLY PRIOR TO 10/23/12 Apache HBase Application Archetypes Lars George | @larsgeorge | Cloudera EMEA Chief Architect | HBase PMC Jonathan Hsieh | @jmhsieh | Cloudera HBase Tech lead | HBase PMC HBaseCon 2014 May 5th , 2014 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 1
About Lars and Jon Lars George • EMEA Chief Architect @Cloudera • Apache HBase PMC • O’Reilly Author of HBase – The Definitive Guide • Contact • lars@cloudera.com • @larsgeorge Jon Hsieh • Tech Lead HBase Team @Cloudera • Apache HBase PMC • Apache Flume founder • Contact: • jon@cloudera.com • @jmhsieh 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 2
About Supporting HBase at Cloudera • Supporting Customers using HBase since 2011 • HBase Training • Professional Services • Team has experience supporting and running HBase since 2009 • 8 committers on staff • 2 HBase book authors • As of Jan 2014, ~20,000 HBase nodes (in aggregate) under management • Information in this presentation is either aggregated customer data or from public sources. 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 3
An Apache HBase Timeline 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 20142008 2009 2010 2011 20132012 Apr’11: CDH3 GA with HBase 0.90.1 May ‘12: HBaseCon 2012 Jun ‘13: HBaseCon 2013 Summer‘11: Messages on HBaseSummer ‘09 StumbleUpon goes production on HBase ~0.20 Nov ‘11: Cassini on HBase Jan ‘13 Phoenix on HBase Summer‘11: Web Crawl Cache 4 Sept’11: HBase TDG published Nov’12: HBase in Action published 2015 May ‘14: HBaseCon 2014 Aug ‘13 Flurry 1k-1k node cluster replication Summer ‘14 HBase v1.0.0 released Jan’14: Cloudera has ~20k Hbase nodes under management
Apache HBase “Nascar” Slide 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 5
Outline • Definitions • Archetypes • The Good • The Bad • The Maybe • Conclusion 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 6
A vocabulary for HBase Archetypes Definitions 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 7
Defining HBase Archetypes • There are a lot of HBase applications • Some successful, some less so • They have common architecture patterns • They have common tradeoffs • Archetypes are common architecture patterns • Common across multiple use-cases • Extracted to be repeatable • Our Goal: Define patterns à la “Gang of Four” (Gamma, Helm, Johnson, Vlissides) 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 8
So you want to use HBase? • What data is being stored? • Entity data • Event data • Why is the data being stored? • Operational use cases • Analytical use cases • How does the data get in and out? • Real time vs. Batch • Random vs. Sequential 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 9
What is being stored? There are primarly two kinds of big data workloads. They have different storage requirements. Entities Events 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 10
Entity Centric Data • Entity data is information about current state • Generally real time reads and writes • Examples: • Accounts • Users • Geolocation points • Click Counts and Metrics • Current Sensors Reading • Scales up with # of Humans and # of Machines/Sensors • Billions of distinct entities 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 11
Event Centric Data • Event centric data are time-series data points recording successive points spaced over time intervals. • Generally real time write, some combination of real time read or batch read • Examples: • Sensor data over time • Historical Stock Ticker data • Historical Metrics • Clicks time-series • Scales up due to finer grained intervals, retention policies, and the passage of time 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 12
Events about Entities • Majority Big Data use cases are dealing with event-based data • |Entities| * |Events| = Big data • When you ask questions, do you hone in on entity first? • When you ask questions, do you hone in on time ranges first? • Your answer will help you determine where and how to store your data. 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 13
Why are you storing the data? • So what kind of questions are you asking the data? • Entity-centric questions • Give me everything about entity e • Give me the most recent event v about entity e • Give me the n most recent events V about entity e • Give me all events V about e between time [t1,t2] • Event and Time-centric questions • Give me an aggregates on each entity between time [t1,t2] • Give me an aggregate on each time interval for entity e • Find events V that match some other given criteria 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 14
How does data get in and out of HBase? HBase Client Put, Incr, Append 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh HBase Client Gets Short scan Full Scan, MapReduce HBase Scanner Bulk Import HBase Client 15 HBase Replication HBase Replication
How does data get in and out of HBase? HBase Client Put, Incr, Append 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh HBase Client Get, Scan Bulk Import HBase Client 16 HBase Replication HBase Replication low latency high throughput Gets Short scan Full Scan, MapReduce HBase Scanner
What system is most efficient? • It is all physics • You have a limited I/O budget • Use all your I/O by parallelizing access and read/write sequentially. • Choose the system and features that reduces I/O in general • Pick the systems best for your workload 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 17 IOPs/s/disk
The physics of Hadoop Storage Systems Workload HBase HDFS Low latency ms, cached mins, MR + seconds, Impala Random Read primary index - index?, small files problem Short Scan sorted + partition Full Scan 0 live table + (MR on snapshots) MR, Hive, Impala Random Write log structured - Not supported Sequential Write hbase overhead bulk load minimal overhead Updates log structured - Not supported 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 18
The physics of Hadoop Storage Systems Workload HBase HDFS Low latency ms, cached mins, MR + seconds, Impala Random Read primary index - index?, small files problem Short Scan sorted + partition Full Scan 0 live table + (MR on snapshots) MR, Hive, Impala Random Write log structured - Not supported Sequential Write hbase overhead bulk load minimal overhead Updates log structured - Not supported 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 19
The physics of Hadoop Storage Systems Workload HBase HDFS Low latency ms, cached mins, MR + seconds, Impala Random Read primary index - index?, small files problem Short Scan sorted + partition Full Scan 0 live table + (MR on snapshots) MR, Hive, Impala Random Write log structured - not supported Sequential Write HBase overhead bulk load minimal overhead Updates log structured - not supported 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 20
The Archetypes HBase Applications 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 21
HBase application use cases • The Good • Simple Entities • Messaging Store • Graph Store • Metrics Store • The Bad • Large Blobs • Naïve RDBMS port • Analytic Archive • The Maybe • Time series DB • Combined workloads 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 22
Archetypes: The Good HBase, you are my soul mate. 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 23
Archetype: Simple Entities • Purely entity data, no relation between entities • Batch or real-time, random writes • Real-time, random reads • Could be a well-done denormalized RDBMS port. • Often from many different sources, with poly-structured data • Schema: • Row per entity • Row key => entity ID, or hash of entity ID • Col qualifier => Property / field, possibly time stamp • Geolocation data • Search index building • Use solr to make text data searchable. 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 24
Simple Entities access pattern HBase Client Put, Incr, Append 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh HBase Client Get, Scan Bulk Import HBase Client 25 HBase Replication low latency high throughput Gets Short scan Full Scan, MapReduce HBase Scanner HBase Replication Solr
Archetype: Messaging Store • Messaging Data: • Realtime Random writes: Emails, SMS, MMS, IM • Realtime random updates: Msg read, starred, moved, deleted • Reading of top-N entries, sorted by time • Records are of varying size • Some time series, but mostly random read/write • Schema: • Row = users/feed/inbox • Row key = UID or UID + time • Column Qualifier = time or conversation id + time. • Use CF’s for indexes. • Examples: • Facebook Messages, Xiaomi Messages • Telco SMS/MMS services • Feeds like tumblr, pinterest 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 26
Facebook Messages - Statistics Source: HBaseCon 2012 - Anshuman Singh 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 27
Messages Access Pattern HBase Client Put, Incr, Append 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh HBase Client Get, Scan Bulk Import HBase Client 28 HBase Replication HBase Replication low latency high throughput Gets Short scan Full Scan, MapReduce HBase Scanner
Archetype: Graph Data • Graph Data: All entities and relations • Batch or realtime, random writes • Batch or realtime, random reads • Its an entity with relation edges • Schema: • Row = Node. • Row key => Node ID. • Col qualifier => Edge ID, or properties:values • Examples: • Web Caches – Yahoo!, Trend Micro • Titan Graph DB with HBase storage backend • Sessionization (financial transactions, clicks streams, network traffic) • Government (connect the bad guy) 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 29
Graph Data Access Pattern HBase Client Put, Incr, Append 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh HBase Client Get, Scan Bulk Import HBase Client 30 HBase Replication HBase Replication low latency high throughput Gets Short scan Full Scan, MapReduce HBase Scanner
Archetype: Metrics • Frequently updated Metrics • Increments • Roll ups generated by MR and bulk loaded to HBase • Poor man’s datacubes • Examples • Campaign Impression/Click counts (Ad tech) • Sensor data (Energy, Manufacturing, Auto) 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 31
Metrics Access Pattern HBase Client Put, Incr, Append 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh HBase Client Get, Scan Bulk Import HBase Client 32 HBase Replication HBase Replication low latency high throughput Gets Short scan Full Scan, MapReduce HBase Scanner
CONFIDENTIAL - RESTRICTED Archetypes: The Bad These are not the droids you are looking for 33
Current HBase weak spots • HBase’s architecture can handle a lot • We make engineering trade offs to optimize for them. • HBase can still do things it is not optimal for. • However, other systems are fundamentally more efficient for some workloads. • We’ve often seen some folks forcing apps into HBase. • If one of these is your only workloads on this data, use another system • If you are in a mixed workload case, some of these become “maybes”. • Just because it is not good today, doesn’t mean it cant be better tomorrow. 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 34
Bad Archetype: Large Blob Store • Saving large objects >3MB per cell • Schema: • Normal entity pattern, but with some columns with large cells. • Examples • Raw photo or video storage in HBase • Large frequently updated structs as a single cell • Problems: • Will get crushed due to write amplification when reoptimizing data for read. (compactions on large unchanging data) • Will crush write pipeline if there are large structs with frequently updated subfields. Cells are atomic, and hbase must rewrite an entire cell. • Some work adding LOB support • This requires new architecture elements 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 35
Bad Archetype: Naïve RDBMS port • A naïve port the RDBMS onto HBase, directly copying the schema. • Schema • Many tables, just like an RDBMS schema. • Row key: primary key or auto-incrementing key, like RDBMS schema • Column qualifiers: field names • Manually do joins, or secondary indexes (not consistent) • Solution: • HBase is not a SQL Database. • No multi-region/multi-table in HBase transactions (yet). • Must to denormalize your schema to use Hbase. 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 36
Large blob store, Naïve RDBMS port access patterns HBase Client Put, Incr, Append 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh HBase Client Get, Scan Bulk Import HBase Client 37 HBase Replication HBase Replication low latency high throughput Gets Short scan Full Scan, MapReduce HBase Scanner
Bad Archetype: Analytic archive • Store purely chronological data, partitioned by time • Real time writes, chronological time as primary index • Column-centric aggregations over all rows. • Bulk reads out, generally for generating periodic reports • Schema • Row key: date+xxx or salt+date+xxx • Column qualifiers: properties with data or counters • Example • Machine logs organized by date. • Full fidelity clickstream 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 38
Bad Archetype: Analytic archive Problems • HBase non-optimal as primary use case. • Will get crushed by frequent full table scans. • Will get crushed by large compactions. • Will get crushed by write-side region hot spotting. • Instead • Store in HDFS; Use Parquet columnar data storage + Impala/Hive • Build rollups in HDFS+MR; store and serve rollups in HBase 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 39
Analytic Archive access patterns HBase Client Put, Incr, Append 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh HBase Client Get, Scan Bulk Import HBase Client 40 HBase Replication HBase Replication low latency high throughput Gets Short scan Full Scan, MapReduce HBase Scanner
And this is crazy | But here’s my data, | serve it, maybe! Archetypes: The Maybe 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 41
The Maybe’s • For some applications, doing it right gets complicated. • These more sophisticated or nuanced cases require considing these questions: • When do you choose HBase vs HDFS storage for time series data? • Are there times where bad archetypes are ok? 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 42
Time Series: in HBase or HDFS? • IO Patterns: • Reads: Collocate related data • Make reads cheap and fast. • Writes: Spread writes out as much as possible • Maximize write throughput • HBase: Tension between these goals • Spreading writes spreads data making reads inefficient • Colocating on write causes hotspots, underutilizes resources by limiting write throughput • HDFS: The sweet spot. • Sequential writes and and sequential read. • Just write more files in date-dirs; physically spreads writes but logically groups data. • Reads for time centric quieries just read files in date-dir 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 43
Time Series data flows • Ingest • Flume or similar direct tool via app • HDFS • Batch queries and generate rollups in Hive/MR • Faster queries in Impala • No user time serving • HBase for recent, HDFS for historical • HBase • Serve individual events • Serve pre-computed aggregates 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 44
Archetype: Entity Time Series • A time series access pattern suitable for HBase • Random write to event data, random read specific event or aggregate data • Generate aggregates via counters, don’t directly compute aggregate on query • HBase is system of record • Schema: • Rowkey: entity-timestamp or hash(entity)-timestamp, possibly with salt added after entity. • Col qualifiers: property • Use custom aggretation to consolidate old data • Use TTL’s to bound and age off old data • Examples: • OpenTSDB does this well for numeric values; Lazily aggregates cells for better performance. • Facebook Insights, ODS 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 45
Entity Time Series access pattern HBase Client Put, Incr, Append 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh HBase Client Get, Scan Bulk Import HBase Client 46 HBase Replication HBase Replication low latency high throughput Gets Short scan Full Scan, MapReduce HBase Scanner Flume Custom App
Archetypes: Hybrid Entity Time Series • Essentially a combo of the Metric Archetype and Entity Time Series Archetype, with bulk loads of rollups via HDFS. • Land data in HDFS and HBase • Keep all data in HDFS for future use • Aggregate in HDFS and write to HBase • HBase can do some aggregates too (counters) • Keep serve-able data in HBase. • Use TTL to discard old values from Hbase. 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 47
Hybrid time series access pattern HBase Client Put, Incr, Append 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh HBase Client Get, Scan Hive or MR: Bulk Import HBase Client 48 HBase Replication HBase Replication low latency high throughput Gets Short scan Full Scan, MapReduce HBase Scanner HDFS Flume
Meta Archetype: Combined workloads • In these cases, the use of HBase depends on workload • Cases where we have multiple workloads styles. • Many cases we want to do multiple things with the same data • primary use case (real time, random access) • secondary use case (analytical) • Pick for your primary, here’s some patterns on how to do your secondary. 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 49
Real time workloads and Analytical access 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh HBase Client Get, Scan 50 poor latency! full scans interfere with latency! high throughput MapReduce HBase Scanner HBase Client Put, Incr, Append Bulk Import HBase Client HBase Replication
Real time workloads and Analytical access 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh HBase Client Get, Scan 51 HBase Replication low latency Isolated from full scans high throughput MapReduce HBase Scanner HBase Client Put, Incr, Append Bulk Import HBase Client HBase Replication high throughput
MR over Table Snapshots (0.98, CDH5.0) • Previously MapReduce jobs over HBase required online full table scan • Take a snapshot and run MR job over snapshot files • Doesn’t use HBase client • Avoid affecting HBase caches • 3-5x perf boost. • Still requires more IOPs than hdfs raw files 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh map map map map map map map map reduce reduce reduce map map map map map map map map reduce reduce reduce snapshot 52
Analytic Archive access pattern HBase Client Put, Incr, Append 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh HBase Client Get, Scan Bulk Import HBase Client 53 HBase Replication HBase Replication low latency high throughput Gets Short scan Full Scan, MapReduce HBase Scanner
Analytic Archive Snapshot access pattern HDFS HBase Client Put, Incr, Append 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh HBase Client Snapshot Scan, MR HBase Scanner Bulk Import HBase Client 54 HBase Replication HBase Replication low latency Higher throughput Table snapshot Gets Short scan
Multitenancy (in progress) • We want to MR for analytics while serving low-latency requests in one cluster. • Performance Isolation • Limit performance impact load on one table has on others. (HBASE- 6721) • Request prioritization and scheduling • Toda default is FIFO • Need to schedule some requests before others (HBASE-10994) 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 55 1 1 2 1 1 3 1 1 1 21 1 31 Delayed by long scan requests Rescheduled so new request get priority Mixed workload Isolated workload
Conclusions 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 56
Big Data Workloads 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 57 Low latency Batch Random Access Full ScanShort Scan HDFS + MR (Hive/pig) HBase HBase + Snapshots -> HDFS + MR HDFS + Impala HBase + MR
Big Data Workloads 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 58 Low latency Batch Random Access Full ScanShort Scan HDFS + MR (Hive/pig) HBase HBase + Snapshots -> HDFS + MR HDFS + Impala HBase + MR Current Metrics Graph data Simple Entities Hybrid Entity Time series + Rollup serving Messages Analytic archive Hybrid Entity Time series + Rollup generation Index building Entity Time series
HBase is evolving to be an Operational Database • Excels at consistent single row centric operations • Dev efforts aimed at using all machine resources efficiently, reducing MTTR, and improving latency predictability. • Projects built on HBase that enable secondary indexing and multi-row transactions • Apache Phoenix (incubating) or Impala provide a SQL skin for simplified application development • Analytic workloads? • Can be done but will be beaten by direct HDFS + MR/Spark/Impala 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 59
Questions? @larsgeorge @jmhsieh 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 60

More Related Content

PDF
HBase Read High Availability Using Timeline-Consistent Region Replicas
byHBaseCon
 
PPTX
Harmonizing Multi-tenant HBase Clusters for Managing Workload Diversity
byHBaseCon
 
PPTX
Content Identification using HBase
byHBaseCon
 
PDF
HBaseCon 2013: Integration of Apache Hive and HBase
byCloudera, Inc.
 
PDF
HBaseCon 2015- HBase @ Flipboard
byMatthew Blair
 
PPT
Real-Time Video Analytics Using Hadoop and HBase (HBaseCon 2013)
bySuman Srinivasan
 
PDF
Data Evolution in HBase
byHBaseCon
 
PPTX
HBaseCon 2013: Using Coprocessors to Index Columns in an Elasticsearch Cluster
byCloudera, Inc.
 
HBase Read High Availability Using Timeline-Consistent Region Replicas
byHBaseCon
 
Harmonizing Multi-tenant HBase Clusters for Managing Workload Diversity
byHBaseCon
 
Content Identification using HBase
byHBaseCon
 
HBaseCon 2013: Integration of Apache Hive and HBase
byCloudera, Inc.
 
HBaseCon 2015- HBase @ Flipboard
byMatthew Blair
 
Real-Time Video Analytics Using Hadoop and HBase (HBaseCon 2013)
bySuman Srinivasan
 
Data Evolution in HBase
byHBaseCon
 
HBaseCon 2013: Using Coprocessors to Index Columns in an Elasticsearch Cluster
byCloudera, Inc.
 

What's hot

PPTX
HBase at Bloomberg: High Availability Needs for the Financial Industry
byHBaseCon
 
PPTX
HBaseCon 2015: HBase and Spark
byHBaseCon
 
PDF
Large-scale Web Apps @ Pinterest
byHBaseCon
 
PPTX
HBaseCon 2012 | Mignify: A Big Data Refinery Built on HBase - Internet Memory...
byCloudera, Inc.
 
PPTX
HBaseCon 2013: Project Valta - A Resource Management Layer over Apache HBase
byCloudera, Inc.
 
PPTX
Apache Spark on Apache HBase: Current and Future
byHBaseCon
 
PPTX
HBaseCon 2012 | Building a Large Search Platform on a Shoestring Budget
byCloudera, Inc.
 
PPTX
HBaseCon 2015: Apache Phoenix - The Evolution of a Relational Database Layer ...
byHBaseCon
 
PPTX
A Graph Service for Global Web Entities Traversal and Reputation Evaluation B...
byHBaseCon
 
PPTX
HBaseCon 2015: Analyzing HBase Data with Apache Hive
byHBaseCon
 
PPTX
HBase in Practice
byDataWorks Summit/Hadoop Summit
 
PPT
HBaseCon 2012 | You’ve got HBase! How AOL Mail Handles Big Data
byCloudera, Inc.
 
PPTX
Taming the Elephant: Efficient and Effective Apache Hadoop Management
byDataWorks Summit/Hadoop Summit
 
PDF
HBaseCon 2015: Graph Processing of Stock Market Order Flow in HBase on AWS
byHBaseCon
 
PPTX
HBaseCon 2013: HBase SEP - Reliable Maintenance of Auxiliary Index Structures
byCloudera, Inc.
 
PPTX
HBase Backups
byHBaseCon
 
PPTX
HBaseCon 2012 | HBase, the Use Case in eBay Cassini
byCloudera, Inc.
 
PPTX
Dancing with the elephant h base1_final
byasterix_smartplatf
 
PPTX
HBaseCon 2013: ETL for Apache HBase
byCloudera, Inc.
 
PPTX
Keynote: The Future of Apache HBase
byHBaseCon
 
HBase at Bloomberg: High Availability Needs for the Financial Industry
byHBaseCon
 
HBaseCon 2015: HBase and Spark
byHBaseCon
 
Large-scale Web Apps @ Pinterest
byHBaseCon
 
HBaseCon 2012 | Mignify: A Big Data Refinery Built on HBase - Internet Memory...
byCloudera, Inc.
 
HBaseCon 2013: Project Valta - A Resource Management Layer over Apache HBase
byCloudera, Inc.
 
Apache Spark on Apache HBase: Current and Future
byHBaseCon
 
HBaseCon 2012 | Building a Large Search Platform on a Shoestring Budget
byCloudera, Inc.
 
HBaseCon 2015: Apache Phoenix - The Evolution of a Relational Database Layer ...
byHBaseCon
 
A Graph Service for Global Web Entities Traversal and Reputation Evaluation B...
byHBaseCon
 
HBaseCon 2015: Analyzing HBase Data with Apache Hive
byHBaseCon
 
HBase in Practice
byDataWorks Summit/Hadoop Summit
 
HBaseCon 2012 | You’ve got HBase! How AOL Mail Handles Big Data
byCloudera, Inc.
 
Taming the Elephant: Efficient and Effective Apache Hadoop Management
byDataWorks Summit/Hadoop Summit
 
HBaseCon 2015: Graph Processing of Stock Market Order Flow in HBase on AWS
byHBaseCon
 
HBaseCon 2013: HBase SEP - Reliable Maintenance of Auxiliary Index Structures
byCloudera, Inc.
 
HBase Backups
byHBaseCon
 
HBaseCon 2012 | HBase, the Use Case in eBay Cassini
byCloudera, Inc.
 
Dancing with the elephant h base1_final
byasterix_smartplatf
 
HBaseCon 2013: ETL for Apache HBase
byCloudera, Inc.
 
Keynote: The Future of Apache HBase
byHBaseCon
 

Viewers also liked

PPTX
HBaseCon 2015: Industrial Internet Case Study using HBase and TSDB
byHBaseCon
 
PDF
HBase for Architects
byNick Dimiduk
 
PDF
Improvements to Apache HBase and Its Applications in Alibaba Search
byHBaseCon
 
PDF
HBase杂谈
byJoseph Pan
 
PDF
HBaseCon 2015: Warcbase - Scaling 'Out' and 'Down' HBase for Web Archiving
byHBaseCon
 
PPTX
HBaseCon 2013: Apache Drill - A Community-driven Initiative to Deliver ANSI S...
byCloudera, Inc.
 
PPTX
Real-time HBase: Lessons from the Cloud
byHBaseCon
 
PPTX
HBaseCon 2015: Blackbird Collections - In-situ Stream Processing in HBase
byHBaseCon
 
PPTX
HBaseCon 2015: HBase Operations in a Flurry
byHBaseCon
 
PPTX
Rolling Out Apache HBase for Mobile Offerings at Visa
byHBaseCon
 
PDF
HBaseCon 2015: Solving HBase Performance Problems with Apache HTrace
byHBaseCon
 
PPTX
Update on OpenTSDB and AsyncHBase
byHBaseCon
 
PDF
HBaseCon 2013: Real-Time Model Scoring in Recommender Systems
byCloudera, Inc.
 
PPTX
Digital Library Collection Management using HBase
byHBaseCon
 
PPTX
HBase Data Modeling and Access Patterns with Kite SDK
byHBaseCon
 
PDF
Apache HBase in the Enterprise Data Hub at Cerner
byHBaseCon
 
PDF
Apache HBase Improvements and Practices at Xiaomi
byHBaseCon
 
PPTX
HBaseCon 2015: HBase @ CyberAgent
byHBaseCon
 
PPTX
HBaseCon 2015: HBase Performance Tuning @ Salesforce
byHBaseCon
 
PPTX
HBaseCon 2013: Full-Text Indexing for Apache HBase
byCloudera, Inc.
 
HBaseCon 2015: Industrial Internet Case Study using HBase and TSDB
byHBaseCon
 
HBase for Architects
byNick Dimiduk
 
Improvements to Apache HBase and Its Applications in Alibaba Search
byHBaseCon
 
HBase杂谈
byJoseph Pan
 
HBaseCon 2015: Warcbase - Scaling 'Out' and 'Down' HBase for Web Archiving
byHBaseCon
 
HBaseCon 2013: Apache Drill - A Community-driven Initiative to Deliver ANSI S...
byCloudera, Inc.
 
Real-time HBase: Lessons from the Cloud
byHBaseCon
 
HBaseCon 2015: Blackbird Collections - In-situ Stream Processing in HBase
byHBaseCon
 
HBaseCon 2015: HBase Operations in a Flurry
byHBaseCon
 
Rolling Out Apache HBase for Mobile Offerings at Visa
byHBaseCon
 
HBaseCon 2015: Solving HBase Performance Problems with Apache HTrace
byHBaseCon
 
Update on OpenTSDB and AsyncHBase
byHBaseCon
 
HBaseCon 2013: Real-Time Model Scoring in Recommender Systems
byCloudera, Inc.
 
Digital Library Collection Management using HBase
byHBaseCon
 
HBase Data Modeling and Access Patterns with Kite SDK
byHBaseCon
 
Apache HBase in the Enterprise Data Hub at Cerner
byHBaseCon
 
Apache HBase Improvements and Practices at Xiaomi
byHBaseCon
 
HBaseCon 2015: HBase @ CyberAgent
byHBaseCon
 
HBaseCon 2015: HBase Performance Tuning @ Salesforce
byHBaseCon
 
HBaseCon 2013: Full-Text Indexing for Apache HBase
byCloudera, Inc.
 

Similar to A Survey of HBase Application Archetypes

PDF
HBase ArcheTypes
byMatteo Bertozzi
 
PPTX
Apache HBase Application Archetypes
byCloudera, Inc.
 
PDF
Intro to HBase - Lars George
byJAX London
 
PDF
Michael stack -the state of apache h base
byhdhappy001
 
PPTX
HBase Tutorial For Beginners | HBase Architecture | HBase Tutorial | Hadoop T...
bySimplilearn
 
PPTX
Hbasepreso 111116185419-phpapp02
byGokuldas Pillai
 
PPTX
Introduction to Apache HBase
byGokuldas Pillai
 
PDF
HBase The Definitive Guide 2 (Early Release) Edition Lars George
bylantoliebe4q
 
PPTX
Hadoop - Apache Hbase
byVibrant Technologies & Computers
 
PDF
Conhecendo o Apache HBase
byFelipe Ferreira
 
PPT
HBASE Overview
bySampath Rachakonda
 
PPTX
hbase hkijhishdfshdfsjkdfsdhfsdhfsjkfjdf
byrethinakumari2
 
PDF
HBase, crazy dances on the elephant back.
byRoman Nikitchenko
 
PPTX
Unit II Hadoop Ecosystem_Updated.pptx
byBhavanaHotchandani
 
PPTX
Hbase.pptx
bymirwais12
 
PDF
Apache Big Data EU 2015 - HBase
byNick Dimiduk
 
PPT
Chicago Data Summit: Apache HBase: An Introduction
byCloudera, Inc.
 
PDF
Nyc hadoop meetup introduction to h base
by智杰 付
 
PPTX
Hadoop and h base in the real world
byJoey Echeverria
 
PDF
Hbase mhug 2015
byJoseph Niemiec
 
HBase ArcheTypes
byMatteo Bertozzi
 
Apache HBase Application Archetypes
byCloudera, Inc.
 
Intro to HBase - Lars George
byJAX London
 
Michael stack -the state of apache h base
byhdhappy001
 
HBase Tutorial For Beginners | HBase Architecture | HBase Tutorial | Hadoop T...
bySimplilearn
 
Hbasepreso 111116185419-phpapp02
byGokuldas Pillai
 
Introduction to Apache HBase
byGokuldas Pillai
 
HBase The Definitive Guide 2 (Early Release) Edition Lars George
bylantoliebe4q
 
Hadoop - Apache Hbase
byVibrant Technologies & Computers
 
Conhecendo o Apache HBase
byFelipe Ferreira
 
HBASE Overview
bySampath Rachakonda
 
hbase hkijhishdfshdfsjkdfsdhfsdhfsjkfjdf
byrethinakumari2
 
HBase, crazy dances on the elephant back.
byRoman Nikitchenko
 
Unit II Hadoop Ecosystem_Updated.pptx
byBhavanaHotchandani
 
Hbase.pptx
bymirwais12
 
Apache Big Data EU 2015 - HBase
byNick Dimiduk
 
Chicago Data Summit: Apache HBase: An Introduction
byCloudera, Inc.
 
Nyc hadoop meetup introduction to h base
by智杰 付
 
Hadoop and h base in the real world
byJoey Echeverria
 
Hbase mhug 2015
byJoseph Niemiec
 

More from HBaseCon

PDF
hbaseconasia2017: Building online HBase cluster of Zhihu based on Kubernetes
byHBaseCon
 
PDF
hbaseconasia2017: HBase on Beam
byHBaseCon
 
PDF
hbaseconasia2017: HBase Disaster Recovery Solution at Huawei
byHBaseCon
 
PDF
hbaseconasia2017: Removable singularity: a story of HBase upgrade in Pinterest
byHBaseCon
 
PDF
hbaseconasia2017: HareQL：快速HBase查詢工具的發展過程
byHBaseCon
 
PDF
hbaseconasia2017: Apache HBase at Netease
byHBaseCon
 
PDF
hbaseconasia2017: HBase在Hulu的使用和实践
byHBaseCon
 
PDF
hbaseconasia2017: 基于HBase的企业级大数据平台
byHBaseCon
 
PDF
hbaseconasia2017: HBase at JD.com
byHBaseCon
 
PDF
hbaseconasia2017: Large scale data near-line loading method and architecture
byHBaseCon
 
PDF
hbaseconasia2017: Ecosystems with HBase and CloudTable service at Huawei
byHBaseCon
 
PDF
hbaseconasia2017: HBase Practice At XiaoMi
byHBaseCon
 
PDF
hbaseconasia2017: hbase-2.0.0
byHBaseCon
 
PDF
HBaseCon2017 Democratizing HBase
byHBaseCon
 
PDF
HBaseCon2017 Removable singularity: a story of HBase upgrade in Pinterest
byHBaseCon
 
PDF
HBaseCon2017 Quanta: Quora's hierarchical counting system on HBase
byHBaseCon
 
PDF
HBaseCon2017 Transactions in HBase
byHBaseCon
 
PDF
HBaseCon2017 Highly-Available HBase
byHBaseCon
 
PDF
HBaseCon2017 Apache HBase at Didi
byHBaseCon
 
PDF
HBaseCon2017 gohbase: Pure Go HBase Client
byHBaseCon
 
hbaseconasia2017: Building online HBase cluster of Zhihu based on Kubernetes
byHBaseCon
 
hbaseconasia2017: HBase on Beam
byHBaseCon
 
hbaseconasia2017: HBase Disaster Recovery Solution at Huawei
byHBaseCon
 
hbaseconasia2017: Removable singularity: a story of HBase upgrade in Pinterest
byHBaseCon
 
hbaseconasia2017: HareQL：快速HBase查詢工具的發展過程
byHBaseCon
 
hbaseconasia2017: Apache HBase at Netease
byHBaseCon
 
hbaseconasia2017: HBase在Hulu的使用和实践
byHBaseCon
 
hbaseconasia2017: 基于HBase的企业级大数据平台
byHBaseCon
 
hbaseconasia2017: HBase at JD.com
byHBaseCon
 
hbaseconasia2017: Large scale data near-line loading method and architecture
byHBaseCon
 
hbaseconasia2017: Ecosystems with HBase and CloudTable service at Huawei
byHBaseCon
 
hbaseconasia2017: HBase Practice At XiaoMi
byHBaseCon
 
hbaseconasia2017: hbase-2.0.0
byHBaseCon
 
HBaseCon2017 Democratizing HBase
byHBaseCon
 
HBaseCon2017 Removable singularity: a story of HBase upgrade in Pinterest
byHBaseCon
 
HBaseCon2017 Quanta: Quora's hierarchical counting system on HBase
byHBaseCon
 
HBaseCon2017 Transactions in HBase
byHBaseCon
 
HBaseCon2017 Highly-Available HBase
byHBaseCon
 
HBaseCon2017 Apache HBase at Didi
byHBaseCon
 
HBaseCon2017 gohbase: Pure Go HBase Client
byHBaseCon
 

Recently uploaded

PPTX
Membershipanywhere - Digital Membership Card - USA , Canada, Australia.pptx
bymembershipanywhere
 
PDF
Microservices-Final !!!!!!!!!!!!!!!!.pdf
bydoquangminh962004
 
PPTX
SAP FICO Training Agenda for new learners
bysasidhar unnagiri
 
PDF
Manual vs Automated Accessibility Testing – What to Choose in 2025.pdf
bykalichargn70th171
 
PDF
Solved First Semester B.C.A. C Programming Question Paper – Jan/Feb 2025
byKuvempu University
 
PDF
Virtual Study Circles Innovative Ways to Collaborate Online.pdf
byExplain Learning
 
PPTX
Modern P&C Insurance Software for Smarter, Faster Operations.pptx
byInsurance Tech Services
 
PDF
Cybersecurity Alert- What Organisations Must Watch Out For This Christmas Fes...
byCybernetic Global Intelligence
 
PDF
Jeremy Millul - An NYU Computer Science Graduate
byJeremy Millul
 
PPTX
application security presentation 2 by harman
byharmanideapad
 
PPTX
NSF Converter Software to Convert NSF to PST, EML, MSG
byNSF Converter Software
 
PPTX
Privacy Preserving Detection of Sensitive Data Exposure (1).pptx
byNikitaMandhan4
 
PDF
BCS-FACS Peter Landin Semantics Seminar - Formal Methods: Whence and Whither?
byJonathan Bowen
 
PDF
Using Agentic RAG for Research - Demonstration of NotebookLM
byRathish Chandra Gatti,Ph.D
 
PDF
DSD-INT 2025 Introduction to Hydrology Suite - Slootjes
byDeltares
 
PDF
DSD-INT 2025 Hydrological Modelling in Society Assumptions, Impacts, and Appl...
byDeltares
 
PPTX
Python-Functions-A-Comprehensive-Overview.pptx
byMuhammad Fahad Bashir
 
PPTX
Computer_Virus_Presentation_With_Slide7.pptx
byHEVISLIVE
 
PPTX
Code Assessment Tools .pptx
byCodexpro
 
PDF
Database Management Systems(DBMS):UNIT-I Introduction to Database(DBMS) BCA S...
byKuvempu University
 
Membershipanywhere - Digital Membership Card - USA , Canada, Australia.pptx
bymembershipanywhere
 
Microservices-Final !!!!!!!!!!!!!!!!.pdf
bydoquangminh962004
 
SAP FICO Training Agenda for new learners
bysasidhar unnagiri
 
Manual vs Automated Accessibility Testing – What to Choose in 2025.pdf
bykalichargn70th171
 
Solved First Semester B.C.A. C Programming Question Paper – Jan/Feb 2025
byKuvempu University
 
Virtual Study Circles Innovative Ways to Collaborate Online.pdf
byExplain Learning
 
Modern P&C Insurance Software for Smarter, Faster Operations.pptx
byInsurance Tech Services
 
Cybersecurity Alert- What Organisations Must Watch Out For This Christmas Fes...
byCybernetic Global Intelligence
 
Jeremy Millul - An NYU Computer Science Graduate
byJeremy Millul
 
application security presentation 2 by harman
byharmanideapad
 
NSF Converter Software to Convert NSF to PST, EML, MSG
byNSF Converter Software
 
Privacy Preserving Detection of Sensitive Data Exposure (1).pptx
byNikitaMandhan4
 
BCS-FACS Peter Landin Semantics Seminar - Formal Methods: Whence and Whither?
byJonathan Bowen
 
Using Agentic RAG for Research - Demonstration of NotebookLM
byRathish Chandra Gatti,Ph.D
 
DSD-INT 2025 Introduction to Hydrology Suite - Slootjes
byDeltares
 
DSD-INT 2025 Hydrological Modelling in Society Assumptions, Impacts, and Appl...
byDeltares
 
Python-Functions-A-Comprehensive-Overview.pptx
byMuhammad Fahad Bashir
 
Computer_Virus_Presentation_With_Slide7.pptx
byHEVISLIVE
 
Code Assessment Tools .pptx
byCodexpro
 
Database Management Systems(DBMS):UNIT-I Introduction to Database(DBMS) BCA S...
byKuvempu University
 

A Survey of HBase Application Archetypes

  • 1.
    Headline Goes Here SpeakerName or Subhead Goes Here DO NOT USE PUBLICLY PRIOR TO 10/23/12 Apache HBase Application Archetypes Lars George | @larsgeorge | Cloudera EMEA Chief Architect | HBase PMC Jonathan Hsieh | @jmhsieh | Cloudera HBase Tech lead | HBase PMC HBaseCon 2014 May 5th , 2014 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 1
  • 2.
    About Lars andJon Lars George • EMEA Chief Architect @Cloudera • Apache HBase PMC • O’Reilly Author of HBase – The Definitive Guide • Contact • lars@cloudera.com • @larsgeorge Jon Hsieh • Tech Lead HBase Team @Cloudera • Apache HBase PMC • Apache Flume founder • Contact: • jon@cloudera.com • @jmhsieh 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 2
  • 3.
    About Supporting HBaseat Cloudera • Supporting Customers using HBase since 2011 • HBase Training • Professional Services • Team has experience supporting and running HBase since 2009 • 8 committers on staff • 2 HBase book authors • As of Jan 2014, ~20,000 HBase nodes (in aggregate) under management • Information in this presentation is either aggregated customer data or from public sources. 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 3
  • 4.
    An Apache HBaseTimeline 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 20142008 2009 2010 2011 20132012 Apr’11: CDH3 GA with HBase 0.90.1 May ‘12: HBaseCon 2012 Jun ‘13: HBaseCon 2013 Summer‘11: Messages on HBaseSummer ‘09 StumbleUpon goes production on HBase ~0.20 Nov ‘11: Cassini on HBase Jan ‘13 Phoenix on HBase Summer‘11: Web Crawl Cache 4 Sept’11: HBase TDG published Nov’12: HBase in Action published 2015 May ‘14: HBaseCon 2014 Aug ‘13 Flurry 1k-1k node cluster replication Summer ‘14 HBase v1.0.0 released Jan’14: Cloudera has ~20k Hbase nodes under management
  • 5.
    Apache HBase “Nascar”Slide 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 5
  • 6.
    Outline • Definitions • Archetypes •The Good • The Bad • The Maybe • Conclusion 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 6
  • 7.
    A vocabulary forHBase Archetypes Definitions 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 7
  • 8.
    Defining HBase Archetypes •There are a lot of HBase applications • Some successful, some less so • They have common architecture patterns • They have common tradeoffs • Archetypes are common architecture patterns • Common across multiple use-cases • Extracted to be repeatable • Our Goal: Define patterns à la “Gang of Four” (Gamma, Helm, Johnson, Vlissides) 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 8
  • 9.
    So you wantto use HBase? • What data is being stored? • Entity data • Event data • Why is the data being stored? • Operational use cases • Analytical use cases • How does the data get in and out? • Real time vs. Batch • Random vs. Sequential 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 9
  • 10.
    What is beingstored? There are primarly two kinds of big data workloads. They have different storage requirements. Entities Events 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 10
  • 11.
    Entity Centric Data •Entity data is information about current state • Generally real time reads and writes • Examples: • Accounts • Users • Geolocation points • Click Counts and Metrics • Current Sensors Reading • Scales up with # of Humans and # of Machines/Sensors • Billions of distinct entities 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 11
  • 12.
    Event Centric Data •Event centric data are time-series data points recording successive points spaced over time intervals. • Generally real time write, some combination of real time read or batch read • Examples: • Sensor data over time • Historical Stock Ticker data • Historical Metrics • Clicks time-series • Scales up due to finer grained intervals, retention policies, and the passage of time 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 12
  • 13.
    Events about Entities •Majority Big Data use cases are dealing with event-based data • |Entities| * |Events| = Big data • When you ask questions, do you hone in on entity first? • When you ask questions, do you hone in on time ranges first? • Your answer will help you determine where and how to store your data. 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 13
  • 14.
    Why are youstoring the data? • So what kind of questions are you asking the data? • Entity-centric questions • Give me everything about entity e • Give me the most recent event v about entity e • Give me the n most recent events V about entity e • Give me all events V about e between time [t1,t2] • Event and Time-centric questions • Give me an aggregates on each entity between time [t1,t2] • Give me an aggregate on each time interval for entity e • Find events V that match some other given criteria 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 14
  • 15.
    How does dataget in and out of HBase? HBase Client Put, Incr, Append 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh HBase Client Gets Short scan Full Scan, MapReduce HBase Scanner Bulk Import HBase Client 15 HBase Replication HBase Replication
  • 16.
    How does dataget in and out of HBase? HBase Client Put, Incr, Append 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh HBase Client Get, Scan Bulk Import HBase Client 16 HBase Replication HBase Replication low latency high throughput Gets Short scan Full Scan, MapReduce HBase Scanner
  • 17.
    What system ismost efficient? • It is all physics • You have a limited I/O budget • Use all your I/O by parallelizing access and read/write sequentially. • Choose the system and features that reduces I/O in general • Pick the systems best for your workload 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 17 IOPs/s/disk
  • 18.
    The physics ofHadoop Storage Systems Workload HBase HDFS Low latency ms, cached mins, MR + seconds, Impala Random Read primary index - index?, small files problem Short Scan sorted + partition Full Scan 0 live table + (MR on snapshots) MR, Hive, Impala Random Write log structured - Not supported Sequential Write hbase overhead bulk load minimal overhead Updates log structured - Not supported 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 18
  • 19.
    The physics ofHadoop Storage Systems Workload HBase HDFS Low latency ms, cached mins, MR + seconds, Impala Random Read primary index - index?, small files problem Short Scan sorted + partition Full Scan 0 live table + (MR on snapshots) MR, Hive, Impala Random Write log structured - Not supported Sequential Write hbase overhead bulk load minimal overhead Updates log structured - Not supported 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 19
  • 20.
    The physics ofHadoop Storage Systems Workload HBase HDFS Low latency ms, cached mins, MR + seconds, Impala Random Read primary index - index?, small files problem Short Scan sorted + partition Full Scan 0 live table + (MR on snapshots) MR, Hive, Impala Random Write log structured - not supported Sequential Write HBase overhead bulk load minimal overhead Updates log structured - not supported 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 20
  • 21.
    The Archetypes HBase Applications 5/5/14HBaseCon 2014; Lars George, Jon Hsieh 21
  • 22.
    HBase application usecases • The Good • Simple Entities • Messaging Store • Graph Store • Metrics Store • The Bad • Large Blobs • Naïve RDBMS port • Analytic Archive • The Maybe • Time series DB • Combined workloads 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 22
  • 23.
    Archetypes: The Good HBase,you are my soul mate. 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 23
  • 24.
    Archetype: Simple Entities •Purely entity data, no relation between entities • Batch or real-time, random writes • Real-time, random reads • Could be a well-done denormalized RDBMS port. • Often from many different sources, with poly-structured data • Schema: • Row per entity • Row key => entity ID, or hash of entity ID • Col qualifier => Property / field, possibly time stamp • Geolocation data • Search index building • Use solr to make text data searchable. 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 24
  • 25.
    Simple Entities accesspattern HBase Client Put, Incr, Append 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh HBase Client Get, Scan Bulk Import HBase Client 25 HBase Replication low latency high throughput Gets Short scan Full Scan, MapReduce HBase Scanner HBase Replication Solr
  • 26.
    Archetype: Messaging Store •Messaging Data: • Realtime Random writes: Emails, SMS, MMS, IM • Realtime random updates: Msg read, starred, moved, deleted • Reading of top-N entries, sorted by time • Records are of varying size • Some time series, but mostly random read/write • Schema: • Row = users/feed/inbox • Row key = UID or UID + time • Column Qualifier = time or conversation id + time. • Use CF’s for indexes. • Examples: • Facebook Messages, Xiaomi Messages • Telco SMS/MMS services • Feeds like tumblr, pinterest 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 26
  • 27.
    Facebook Messages -Statistics Source: HBaseCon 2012 - Anshuman Singh 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 27
  • 28.
    Messages Access Pattern HBaseClient Put, Incr, Append 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh HBase Client Get, Scan Bulk Import HBase Client 28 HBase Replication HBase Replication low latency high throughput Gets Short scan Full Scan, MapReduce HBase Scanner
  • 29.
    Archetype: Graph Data •Graph Data: All entities and relations • Batch or realtime, random writes • Batch or realtime, random reads • Its an entity with relation edges • Schema: • Row = Node. • Row key => Node ID. • Col qualifier => Edge ID, or properties:values • Examples: • Web Caches – Yahoo!, Trend Micro • Titan Graph DB with HBase storage backend • Sessionization (financial transactions, clicks streams, network traffic) • Government (connect the bad guy) 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 29
  • 30.
    Graph Data AccessPattern HBase Client Put, Incr, Append 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh HBase Client Get, Scan Bulk Import HBase Client 30 HBase Replication HBase Replication low latency high throughput Gets Short scan Full Scan, MapReduce HBase Scanner
  • 31.
    Archetype: Metrics • Frequentlyupdated Metrics • Increments • Roll ups generated by MR and bulk loaded to HBase • Poor man’s datacubes • Examples • Campaign Impression/Click counts (Ad tech) • Sensor data (Energy, Manufacturing, Auto) 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 31
  • 32.
    Metrics Access Pattern HBaseClient Put, Incr, Append 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh HBase Client Get, Scan Bulk Import HBase Client 32 HBase Replication HBase Replication low latency high throughput Gets Short scan Full Scan, MapReduce HBase Scanner
  • 33.
    CONFIDENTIAL - RESTRICTED Archetypes:The Bad These are not the droids you are looking for 33
  • 34.
    Current HBase weakspots • HBase’s architecture can handle a lot • We make engineering trade offs to optimize for them. • HBase can still do things it is not optimal for. • However, other systems are fundamentally more efficient for some workloads. • We’ve often seen some folks forcing apps into HBase. • If one of these is your only workloads on this data, use another system • If you are in a mixed workload case, some of these become “maybes”. • Just because it is not good today, doesn’t mean it cant be better tomorrow. 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 34
  • 35.
    Bad Archetype: LargeBlob Store • Saving large objects >3MB per cell • Schema: • Normal entity pattern, but with some columns with large cells. • Examples • Raw photo or video storage in HBase • Large frequently updated structs as a single cell • Problems: • Will get crushed due to write amplification when reoptimizing data for read. (compactions on large unchanging data) • Will crush write pipeline if there are large structs with frequently updated subfields. Cells are atomic, and hbase must rewrite an entire cell. • Some work adding LOB support • This requires new architecture elements 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 35
  • 36.
    Bad Archetype: NaïveRDBMS port • A naïve port the RDBMS onto HBase, directly copying the schema. • Schema • Many tables, just like an RDBMS schema. • Row key: primary key or auto-incrementing key, like RDBMS schema • Column qualifiers: field names • Manually do joins, or secondary indexes (not consistent) • Solution: • HBase is not a SQL Database. • No multi-region/multi-table in HBase transactions (yet). • Must to denormalize your schema to use Hbase. 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 36
  • 37.
    Large blob store,Naïve RDBMS port access patterns HBase Client Put, Incr, Append 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh HBase Client Get, Scan Bulk Import HBase Client 37 HBase Replication HBase Replication low latency high throughput Gets Short scan Full Scan, MapReduce HBase Scanner
  • 38.
    Bad Archetype: Analyticarchive • Store purely chronological data, partitioned by time • Real time writes, chronological time as primary index • Column-centric aggregations over all rows. • Bulk reads out, generally for generating periodic reports • Schema • Row key: date+xxx or salt+date+xxx • Column qualifiers: properties with data or counters • Example • Machine logs organized by date. • Full fidelity clickstream 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 38
  • 39.
    Bad Archetype: Analyticarchive Problems • HBase non-optimal as primary use case. • Will get crushed by frequent full table scans. • Will get crushed by large compactions. • Will get crushed by write-side region hot spotting. • Instead • Store in HDFS; Use Parquet columnar data storage + Impala/Hive • Build rollups in HDFS+MR; store and serve rollups in HBase 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 39
  • 40.
    Analytic Archive accesspatterns HBase Client Put, Incr, Append 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh HBase Client Get, Scan Bulk Import HBase Client 40 HBase Replication HBase Replication low latency high throughput Gets Short scan Full Scan, MapReduce HBase Scanner
  • 41.
    And this iscrazy | But here’s my data, | serve it, maybe! Archetypes: The Maybe 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 41
  • 42.
    The Maybe’s • Forsome applications, doing it right gets complicated. • These more sophisticated or nuanced cases require considing these questions: • When do you choose HBase vs HDFS storage for time series data? • Are there times where bad archetypes are ok? 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 42
  • 43.
    Time Series: inHBase or HDFS? • IO Patterns: • Reads: Collocate related data • Make reads cheap and fast. • Writes: Spread writes out as much as possible • Maximize write throughput • HBase: Tension between these goals • Spreading writes spreads data making reads inefficient • Colocating on write causes hotspots, underutilizes resources by limiting write throughput • HDFS: The sweet spot. • Sequential writes and and sequential read. • Just write more files in date-dirs; physically spreads writes but logically groups data. • Reads for time centric quieries just read files in date-dir 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 43
  • 44.
    Time Series dataflows • Ingest • Flume or similar direct tool via app • HDFS • Batch queries and generate rollups in Hive/MR • Faster queries in Impala • No user time serving • HBase for recent, HDFS for historical • HBase • Serve individual events • Serve pre-computed aggregates 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 44
  • 45.
    Archetype: Entity TimeSeries • A time series access pattern suitable for HBase • Random write to event data, random read specific event or aggregate data • Generate aggregates via counters, don’t directly compute aggregate on query • HBase is system of record • Schema: • Rowkey: entity-timestamp or hash(entity)-timestamp, possibly with salt added after entity. • Col qualifiers: property • Use custom aggretation to consolidate old data • Use TTL’s to bound and age off old data • Examples: • OpenTSDB does this well for numeric values; Lazily aggregates cells for better performance. • Facebook Insights, ODS 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 45
  • 46.
    Entity Time Seriesaccess pattern HBase Client Put, Incr, Append 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh HBase Client Get, Scan Bulk Import HBase Client 46 HBase Replication HBase Replication low latency high throughput Gets Short scan Full Scan, MapReduce HBase Scanner Flume Custom App
  • 47.
    Archetypes: Hybrid EntityTime Series • Essentially a combo of the Metric Archetype and Entity Time Series Archetype, with bulk loads of rollups via HDFS. • Land data in HDFS and HBase • Keep all data in HDFS for future use • Aggregate in HDFS and write to HBase • HBase can do some aggregates too (counters) • Keep serve-able data in HBase. • Use TTL to discard old values from Hbase. 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 47
  • 48.
    Hybrid time seriesaccess pattern HBase Client Put, Incr, Append 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh HBase Client Get, Scan Hive or MR: Bulk Import HBase Client 48 HBase Replication HBase Replication low latency high throughput Gets Short scan Full Scan, MapReduce HBase Scanner HDFS Flume
  • 49.
    Meta Archetype: Combinedworkloads • In these cases, the use of HBase depends on workload • Cases where we have multiple workloads styles. • Many cases we want to do multiple things with the same data • primary use case (real time, random access) • secondary use case (analytical) • Pick for your primary, here’s some patterns on how to do your secondary. 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 49
  • 50.
    Real time workloadsand Analytical access 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh HBase Client Get, Scan 50 poor latency! full scans interfere with latency! high throughput MapReduce HBase Scanner HBase Client Put, Incr, Append Bulk Import HBase Client HBase Replication
  • 51.
    Real time workloadsand Analytical access 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh HBase Client Get, Scan 51 HBase Replication low latency Isolated from full scans high throughput MapReduce HBase Scanner HBase Client Put, Incr, Append Bulk Import HBase Client HBase Replication high throughput
  • 52.
    MR over TableSnapshots (0.98, CDH5.0) • Previously MapReduce jobs over HBase required online full table scan • Take a snapshot and run MR job over snapshot files • Doesn’t use HBase client • Avoid affecting HBase caches • 3-5x perf boost. • Still requires more IOPs than hdfs raw files 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh map map map map map map map map reduce reduce reduce map map map map map map map map reduce reduce reduce snapshot 52
  • 53.
    Analytic Archive accesspattern HBase Client Put, Incr, Append 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh HBase Client Get, Scan Bulk Import HBase Client 53 HBase Replication HBase Replication low latency high throughput Gets Short scan Full Scan, MapReduce HBase Scanner
  • 54.
    Analytic Archive Snapshotaccess pattern HDFS HBase Client Put, Incr, Append 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh HBase Client Snapshot Scan, MR HBase Scanner Bulk Import HBase Client 54 HBase Replication HBase Replication low latency Higher throughput Table snapshot Gets Short scan
  • 55.
    Multitenancy (in progress) •We want to MR for analytics while serving low-latency requests in one cluster. • Performance Isolation • Limit performance impact load on one table has on others. (HBASE- 6721) • Request prioritization and scheduling • Toda default is FIFO • Need to schedule some requests before others (HBASE-10994) 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 55 1 1 2 1 1 3 1 1 1 21 1 31 Delayed by long scan requests Rescheduled so new request get priority Mixed workload Isolated workload
  • 56.
    Conclusions 5/5/14 HBaseCon 2014;Lars George, Jon Hsieh 56
  • 57.
    Big Data Workloads 5/5/14HBaseCon 2014; Lars George, Jon Hsieh 57 Low latency Batch Random Access Full ScanShort Scan HDFS + MR (Hive/pig) HBase HBase + Snapshots -> HDFS + MR HDFS + Impala HBase + MR
  • 58.
    Big Data Workloads 5/5/14HBaseCon 2014; Lars George, Jon Hsieh 58 Low latency Batch Random Access Full ScanShort Scan HDFS + MR (Hive/pig) HBase HBase + Snapshots -> HDFS + MR HDFS + Impala HBase + MR Current Metrics Graph data Simple Entities Hybrid Entity Time series + Rollup serving Messages Analytic archive Hybrid Entity Time series + Rollup generation Index building Entity Time series
  • 59.
    HBase is evolvingto be an Operational Database • Excels at consistent single row centric operations • Dev efforts aimed at using all machine resources efficiently, reducing MTTR, and improving latency predictability. • Projects built on HBase that enable secondary indexing and multi-row transactions • Apache Phoenix (incubating) or Impala provide a SQL skin for simplified application development • Analytic workloads? • Can be done but will be beaten by direct HDFS + MR/Spark/Impala 5/5/14 HBaseCon 2014; Lars George, Jon Hsieh 59
  • 60.