LOGO Discussion Zhang Gang 2012/11/8. Discussion Progress on HBase 1 Cassandra or HBase 2.

LOGO

Discussion Zhang Gang

2012/11/8

Discussion

Progress on HBase1

Cassandra or HBase2

HBase Sechma Design

HBase reference guide How to design a good HBase schema.

– row key– column family

HBase Sechma Design

row key– monotonically increasing keys or timeseries

keys may cause a pile-up on a single region.– randomize the input records to not be in

sorted order can mitigate the situation. So its best to avoid using a timestamp or a sequence as the row key.

– at present, I use the startTime(a timestamp) as the row key, in future I will explore if there has a better replacement.

HBase Sechma Design

column famliy:– I was wrong about the schema with two

column families. – HBase currently does not do well with

anything above two or three column families.– Try to make do with one column family if you

can in your schemas.– If you have thousands or even millions

column, you can consider have more than one column family. We only have 21 columns, so one is enough and the best choice.

HBase Sechma Design

Optimization(minimize row and column sizes)– in HBase, values are always as a cell value that

accompanied by its row, column name, and timestamp. So if row and column name is long, it will waste a large size.(see behind)

– column family: keep the name as short as possible.

– row key length: keep them as short as is reasonable such that they can still be useful for required data access.

Sqoop

Have successfully configured the sqoop in my PC. On farm, have a Exception--”access denied for user ‘zhang’, but it seems successfully transfer the data.

Command：– Sqoop import –connect jdbc:

Sqoop

sqoop on my PC:– test: 81,280 records, 45.1613s– test: 215,500 records, 73.2617s– test: 1,539,763 records,310s– then:35,427,339 records, 1235060s/about

3.43h– the HBase table size: about 35G, compare

mysql table(5G), the size is bigger. So design a good schema is very necessary.

Sqoop

sqoop on the farm:– two exceptions:

– then found access denied– import: 35,427,339 records,5120s/about

1.39h– hbase-name:’hb_type_job’– row-key: ’startTime’– column-family: ’d’

s

Sqoop

Cassandra or HBase

Cassandra or HBase

review our requirement:– big data: now 5G, increases 1.5 GB per year,

not very big.– high scalability: we want the database we

choice has a better scalability.(many candidates have the feature.

– write/read: we read more than we write.(One of the reasons we choose HBase before)

Cassandra or HBase

Written in: Java Main point: Best of BigTable and Dynamo Tunable trade-offs for distribution and replication (N, R, W) Querying by column, range of keys BigTable-like features: columns, column families Has secondary indices Writes are much faster than reads (!) Map/reduce possible with Apache Hadoop All nodes are similar, as opposed to Hadoop/Hbase Gossip protocol, multi data center, no single point of failure

Cassandra or HBase

C has only one type of nodes, all nodes are similar . H consists of several different types of nodes (Muster/RegionServer).

H must deployed over the HDFS, compare this C is much more simple

Data consistency of C is tunable(N,W,R). H better support map/reduce H provides the developer with row locking facilities whereas

Cassandra can not. C just use timestamp. C has better I/O performance and better scalability but not

good at range scan. CAP:C focus on AC and H focus on CP H has an SQL compatibility interface(Hive),so H support SQL

Cassandra or HBase

The structure of C is simple ,deploy and maintenance is simple, compare C(save money, save time) ,H is much more complex deploy or maintenance.

H maybe more suitable for data warehousing, and large scale data processing and analysis. And C being more suitable for real time transaction processing and the serving of interactive data.

Cassandra or HBase

How do I incorporate my logo to a slide that will apply to all the other slides?– bb

Aa– bb

Aa– On

Cassandra or HBase

the possibility we start to explore Cassandra– more simple than Hadoop HBase.– written by Java.(same as HBase)– pycassa:It is a python client library for Apache

Cassandra.

problem: seem doesn’t have a ready-made tool for transfer the data from mysql to Cassandra.

LOGO

Your Company Slogan