Big Data .. Are you ready for the next wave?

Big Data…Are you ready for the next wave?

MAHMOUD SABRY

Agenda

The Next Wave of Computing

Information, Data & Knowledge

Data warehouse vs. Database

Row-oriented vs. Column-oriented DB

Big Data Era

HP-Vertica

HP-Vertica Bulk Upload Example

The next wave of computing

Cloud + Big Data Analytics, Mobile & Social leads to what IBM calls “The Era of Smart.”

Data, Information & Knowledge (1/3)

Data are raw facts, and have no meaning on their own

Example

Yes, Yes, No, Yes, No, Yes, No, Yes

Data, Information & Knowledge (2/3)

Information is data that have meaning in a context

Example

Yes, Yes, No, Yes, No, Yes, No, Yes, No, Yes, YesRaw Data

ContextResponses to the market

research question – “Would you buy brand x at price y?”

Information ???

Processing

Data, Information & Knowledge (3/3)

Knowledge

understanding of someone or something

acquired through experience or learning

Example

Based on last collected information, A Marketing Manager could use this information to decide whether or not to raise or lower price.

Data Warehouse vs. Database (1/2)

Feature DWH DB

Data Stored  It usually stores the Historical data whose accuracy is maintained over time.

It mainly stores the Current data which always guaranteed to be up-to-date.

Characteristic It is based on Informational Processing.

It is based on Operational Processing.

Function It is used for long-term informational requirements and decision support.

It is used for day-to-day operations.

Focus The focus is on “Information OUT” The focus is on “Data IN”

Number of records accessed

A bunch of millions of records. A few tens of records.

Data Warehouse vs. Database (2/2)

Feature DWH DB

Access It mostly use the read access for the stored data.

The most frequent type of access type is read/write.

Orientation Based on Analysis. Based on Transaction.

Common users

Analysts DBAs

KPI Query throughput Transaction throughput

Unit of work Complex queries Short and simple transactions

Row-based vs. Column-based DB (1/2)

Feature Row-oriented DB Column-oriented DB

Mechanism stores table records in sequence or rows.

stores table records in sequence or columns.

Retrieval of a few columns

Slower Faster

Retrieval of a single record

Faster Slower

Row-based vs. Column-based DB (2/2)

Feature Row-oriented DB Column-oriented DB

Insertion/Updating of single new record

Faster Slower

Aggregation of Single Column, e.g. sum(price)

Slower Faster

Compression Higher. As stores similar data together

It depends!

Examples Oracle OLTPMS SQL Server up to 2008

HP-VerticaSybase IQMS SQL Server 2012

Big Data Era – Famous quote

From the dawn of civilization until 2003, humankind generated five exabytes of data. Now we produce five exabytes every two days…and the pace is accelerating.

Eric Schmidt,Executive Chairman, Google

Big Data – Definition

The basic idea behind the phrase 'Big Data' is that everything we do is increasingly leaving a digital trace (or data), which we (and others) can use and analyze.

Big Data therefore refers to our ability to make use of the ever-increasing volumes of data.

Big Data – Characteristics (6 V’s)

Volume Terabytes, Distributed, Tables & Files

Velocity Real-time processing/Streams

Variety Structured, Semi-structured & Un-structured data

Value Correlations & Statistical analysis

Veracity Accountability, Trust, Origin & Reputation

Variability Changing Data/model

Big Data – What is Hadoop ?

Apache™ Hadoop® is an open source platform that enables distributed processing of large data sets across clusters of servers.

It is designed to scale up from a single server to thousands of machines, with very high degree of fault tolerance. 

Big Data – HL Archit. Of Hadoop

A programming model for large scale data processing.

A distributed file-system that stores data on Clustered machines.

Schedules map or reduce jobs to task trackers with an awareness of the data location.

keeps the directory tree of all files in the file system, and tracks where across the cluster the file data is kept.

Big Data – MapReduce, How it works ?

Map: Filter & sort of data Reduce: Summarize & aggregate of data

HP-Vertica

An analytic database management platform. Founded in 2005.

Helps you monetize all of your data in real-time and at massive scale.

Queries run 50-1,000x faster.

Store 10-30x more data per server.

Openness and simplicity (use any BI/ETL tools, Hadoop, etc.)

Suited for “Structured” data only.

HP Vertica Flex Zone: built on the HP Vertica core, enables load and analyze structured and semi-structured data, such as social media, sensor, log files, and machine data.

HP-Vertica Features (1/6)

Column Orientation

Vertica organizes data for each column

Each column is stored separately on disk

Only reads the columns needed to answer the query

Significant reduction of disk I/O

HP-Vertica Features (2/6)

Advanced Compression

Vertica replaces slower disk I/O with faster CPU cycles to encode data elements into a more compact form

and query them.

Vertica’s innovative query engine operates directly on compressed data, meaning that it can actually require fewer CPU operations to process the compressed version of a table.

HP-Vertica Features (3/6)

High Availability

RAID-like functionality within database

If a node fails, a copy is available on one of the surviving nodes

Always-on Queries and Loads

System continues to load and query when nodes are down

Automatically recovers missing data by querying other nodes

HP-Vertica Features (4/6)

Automatic Database Designer (DBD)

Recommends a physical DB design that provides the best performance for the user's workload

Analyzes your logical schema, sample data, and sample queries

Minimizes DBA tuning

Run anytime for additional optimization, without stopping the database

HP-Vertica Features (5/6)

Massively Parallel Processing (MPP)

Parallel design leverages data projections to enable distributed storage and workload

Active redundancy

Automatic replication, failover and recovery

HP-Vertica Features (6/6)

Native SQL and Application Integration

Standard SQL Interface

Simple integration with Hadoop and existing BI and ETL tools

Supports SQL, ODBC, JDBC and majority ETL and BI reporting products

HP-Vertica Bulk Upload Example (1/3)

HP Vertica 7.0, Single instance on VMware workstation v.11

Installed on Windows 8 machine HP-EliteBook 8440p, 8GB RAM

HP Vertica virtual machine has below specs:

Memory: 4GB

Disk: 16 GB

Processor: Single Processor Intel(R) Core(TM) i7 CPU M 620 @ 2.67GHz

HP-Vertica Bulk Upload Example (2/3)

Scenario Details

Upload a table with “|” delimited columns from a local text file

Table file size is: 1.30099 GB

No. of rows is: 10252864

No. of columns Is: 21

Table is partitioned into 10 partitions by using “customer_age“ column

Only one projection (the default “<table_name>_Super” )

HP-Vertica Bulk Upload Example (3/3)

Scenario Actions and Results

Upload by using “DIRECT” option to load directly to disk (ROS: Read Optimized Storage)

COPY public.myTable from '/opt/vertica/examples/VMart_Schema/myTable.tbl' DIRECT;

Time taken is more than 4 minutes

Upload without using “DIRECT” option to load to memory(WOS: Write Optimized Storage)

COPY public.myTable from '/opt/vertica/examples/VMart_Schema/myTable.tbl';

Time taken is about 2.3 minutes

Thanks!