A Foundation for Success in the Information Economy

Twitter Tag: #briefr

The Briefing Room

[email protected]


The Briefing Room

!  Reveal the essential characteristics of enterprise software, good and bad

!  Provide a forum for detailed analysis of today’s innovative technologies

!  Give vendors a chance to explain their product to savvy analysts

!  Allow audience members to pose serious questions... and get answers!


The Briefing Room

November: Cloud

December: Innovators

January: Big Data

February: Performance

March: Integration


The Briefing Room

!  Most organizations and companies rely heavily on databases and database management systems for their operations and processes.

!  The landslide of complex data has not diminished the expectation for reliable performance of and immediate access to database systems.

!  Further, the global drive for 24/7 mission-critical computing

has created challenges in the areas of fault tolerance and scalability, meaning database technology must not only be available on demand, but it must scale on demand, and do so without fail.


The Briefing Room

Robin Bloor is Chief Analyst at The Bloor Group

[email protected]


The Briefing Room

!   Established in 1939, HP specializes in developing and manufacturing computing, data storage and networking hardware; designing software; and delivering services.

!   Though it holds the highest market share of global PC sales (17.2% last year), it has also been building a formidable collection of commercial information management solutions.

!   A key offering is its HP Integrity NonStop, a platform aimed at mission-critical customers that includes an integrated stack of hardware, OS, database, software and applications.


The Briefing Room

Ajaya Gummadi is the Database Product Manager with the HP NonStop Enterprise Division. In this role, she is responsible for setting the database product strategy for the NonStop Business Unit. She engages customers to understand their business and technology requirements, and working with Partners has developed a database ecosystem for the NonStop platform. She works closely with R&D on prioritizing and delivering database innovations that enable customers to create scalable and always available NonStop SQL applications. Working closely with worldwide Sales teams, she evangelizes new product messaging and drives product marketing execution priorities. Ajaya is a Computer Science graduate from BITS Pilani, India and received an EMBA from Pepperdine University’s Graziadio School of Business Management.

© Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice.

A Foundation for Success in the Information Economy NonStop SQL: Mission-critical database

Ajaya Gummadi HP NonStop Database Product Management [email protected] The Briefing Room, October 30, 2012

© Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice. 10

In today’s need-it-now world…

when is it okay for your business to be unavailable to your customers? Never.


When application availability is vital

Your mission-critical experience matters

SECONDS Downtime specified in

System issues

PREDICTED

MINIMAL Software updates

CORRECTED

ZERO Unplanned downtime

Data integrity

NEVER COMPROMISED


Average business revenue lost per hour of downtime (US$)

Business failure brings a high cost

Retail $1,999,872

Healthcare $4,223,520

Manufacturing $10,432,800

Communications $15,120,000

Average $9,700,000

Source: © 2009 HP internal testing and development over two-year period and other competitive materials, including IDC “Cost of Downtime Tool” developed for HP

Financial $16,833,600


•  Diverse data •  Large volumes •  Extreme Velocity •  Instant Access •  Operational and Predictive Analytics

Market Forces Affect Database Industry


Performance •  Need better response times •  Continuous data availability requirements

Scalability •  More data, More Users

Capability •  Real-time, recommendations, mining

Cost and complexity •  Out-of-the-box execution efficiencies,

consolidate workloads, reduce costs, improve SLAs

Market Forces Affect Database Choices


•  Continuously available data architecture

•  Economical and highly scalable architecture

•  Ability to handle high volumes of data

•  Ability to handle variety of data •  Ability to handle velocity of data

Market Requirements Affect Database Architecture


Key NonStop SQL Attributes Integrated hardware and software stack

Out-of-the-box cluster aware

Mixed Workload – Out-of-the-box

Virtualization – data and workload execution

Support for ANSI and Connectivity standards

Support for Oracle syntax

Online manageability

Leverages and fully aligned with NonStop

server’s takeover and MPP architecture


Architected for availability, scalability, and performance

•  Shared-nothing MPP

•  Data virtualization

•  Parallel query execution

•  Out-of-the-box Mixed workload &

transactional processing

•  Cluster aware

•  Unrivaled availability


Scalable to 4,096 processors

Data virtualization …

Transparent software virtualization …

ODBC/JDBC connections

Query processing

Data Access Manager (DAM)

Administration and management

Scalable infrastructure virtualization Processor n

MXCS MXCS .....

ESP ESP .....

DAM DAM .....

NonStop System

Pro

cess

or


Transparent software virtualization Query Processing – parallelizing the work

Query divided into operators Nested, merge, hash joins; unions; partial & full aggregations; sorts; input/output operations (scan, update, delete, insert)

Operator parallelism

Partitioned parallelism

Pipeline parallelism

Connect

Join

Scan Group by

Scan

40

Data-flow, scheduler-driven Parallelism throughout

Varying degrees of parallelism

Scan

Scan

Join

Group by

– Operator executed by an DAM, ESP, or Connect process

– ESPs started for desired degree of parallelism

30 20


Unrivaled availability Elimination of unplanned downtime

Reliable and failure resilient hardware

35+ years of proven HP NonStop system

engineering

Continuously available, in spite of any

single point hardware or software failure

Survives many multi-component failures

Automatically rebalances after component

repair and reintegration

Patented fault-tolerant software “process-

pair” technology


…

PS PS CS CS

X Fabric Y Fabric

P01 P14

M01 M14

Node 15 Node 16

…

Node 1 Node 2

Query or Application

Redirected SQL operation

DAM P10

Cac

he

DAM

B10

Cac

he

Unrivaled availability Fault-tolerant process pairing Patented HP “process pairing” technology

Automatic checkpoint of volatile SQL operations

Inherently resilient to transient software failures

Takeover as opposed to failover

Failure does not interrupt the database availability

No need to restart the database

No database recovery operations

Checkpoint


Customers realize scalability and availability benefits of NonStop SQL

•  A NonStop customer manages several PBs of database with 2 DBAs, pushing 100,000+ tps and has not had any downtime since 1995

•  A securities company migrates to NonStop SQL for its superior availability; cost of downtime was $3M/hour

•  Another NonStop customer manages hundreds of Terabytes of data and has had no outage since going live several years ago

NonStop SQL handles critical business needs


Customers realize mixed workload benefits of NonStop SQL

•  Customer manages 150++TB of NonStop SQL database

•  Drives mixed-workload consisting of 39,000 ingests/second concurrently with >5000 ad-hoc and OLAP queries, and database maintenance activities concurrently

•  Mixed workload capabilities are available out-of-the-box

•  No need of application partitioning and multi-tier complex architectures to workaround lack of these capabilities



Customers realize availability benefits of NS SQL


•  Customer moves to NonStop SQL for its superior availability and TCO

•  Objective was to manage the transactions with no unplanned downtime

•  Performing at 2000 tps, driving 20,000 sql statements from 10,000 concurrent users over 2000 JDBC connections


Customers realize modern and standardization benefits of NS SQL


•  A customer selects NonStop SQL/MX to consolidate distributed

databases into an ODS

•  Application hosted in cloud and accesses NS SQL/MX

•  Customer selected NonStop SQL/MX for its modern and standard

software interfaces

•  Customer relies on NonStop scalability, availability and TCO


–  NonStop SQL delivers out-of-the-box cluster awareness and

management, lowering operational costs

–  No add-on cluster software download and configuration

– Adding resources to a cluster is done online in simple steps, complexity taken away

–  It takes only 19 steps from receving media to having a NonStop SQL database instance up and runniing

– NonStop SQL deploys as a single clustered database image across the entire cluster

NonStop SQL handles critical business needs Customers enjoy clustering and lack of complexity benefits


–  NonStop SQL customers can start small and scale-out flexibly

•  Customers can start small with a 2-core or 4-core NonStop database

server

•  And scale to thousands of cores

•  Customers scale user data from 146 GB to Petabytes

•  There are no prescriptive constraints on how to scale the server in

response to growth in business

Customers enjoy flexible configurations



Low Cost of Acquisition means more for saving money with NonStop SQL

•  All DBA productivity tools are included with the base SQL license with

no additional costs

•  No additional Partitioning Software licenses required

•  Diagnostic, Tuning, Management packs are all included in the base

license

•  NonStop SQL deploys and is managed as a single clustered database

image

•  NonStop has fewer moving parts and less complexity, leading to lower

operational costs

Optimize your database environment


NonStop SQL has an edge in today’s information economy Strong value prop Stronger Proof Points

NonStop SQL is positioned for a takeoff Strong roadmap Investing for the future

29

Key takeaways

© Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice.

Thank you

[email protected]


The Briefing Room

The Bloor Group

The Bloor Group

The Bloor Group

The Bloor Group

The “Big Data” Trend

q  Corporate data volumes grow at about 55% per annum

q  VLDB volumes grow at about 55% per annum

q  This is exponential q  Data has been growing

at this rate for at least 20 years

q  As such there is nothing new about big data other than the current data volumes - which follow a well established trend

The Bloor Group

Horses For Courses

•  RDBMS •  Object DBMS •  Column stores •  Big Table stores •  NoSQL DBMS •  Mixed Workload

DBMS

•  Traditional OLTP or DW •  Objects and OLTP •  Scale Out Analytics •  Log file and sensor data •  Documents and Objects •  Large scale OLTP and

DW + analytics

The Bloor Group

The Data Flow Issue

The Bloor Group

The Advantages

q  Fewer databases can mean fewer points of failure

q  It can mean lower DBA overhead

q  It can mean simpler recovery

q  It is very likely to mean lower latency for BI

applications

q  It can mean lower software costs

q  These advantages can multiply in a mixed workload

environment

The Bloor Group

Questions

1.   Aside from the NonStop architecture what do you believe are the “technical uniques” of NonStop SQL?

2.   What mixed workloads are possible with NonStop SQL?

3.   What areas of application do you regard as its sweet spots?

4.   What is the largest NonStop SQL database (by data volume) currently in use? What is the largest that has a mixed workload?

The Bloor Group

Questions

5.   Where does NonStop SQL sit in relation to HP’s Vertica database?

6.   How difficult is it to use (in other words, what are the labor/DBA overheads compared to a traditional RDBMS)?

7.   What is HP’s strategy in respect to NonStop SQL and Hadoop?

8.   Which database products do you tend to find yourself in competition with?


The Briefing Room


The Briefing Room

November: Cloud

December: Innovators

January: Big Data

2013 Editorial Calendar (www.insideanalysis.com)


The Briefing Room

A Foundation for Success in the Information Economy

Technology

database systems

briefrthe briefing room

database innovations

database technology

database ecosystem

database product strategyfor

database management

data storage