The future of scaling forrester research - GigaSpaces Road Show 2011

44
© 2011 Forrester Research, Inc. Reproduction Prohibited 1

description

 

Transcript of The future of scaling forrester research - GigaSpaces Road Show 2011

Page 1: The future of scaling forrester research - GigaSpaces Road Show 2011

© 2011 Forrester Research, Inc. Reproduction Prohibited1

Page 2: The future of scaling forrester research - GigaSpaces Road Show 2011

© 2011 Forrester Research, Inc. Reproduction Prohibited2 © 2009 Forrester Research, Inc. Reproduction Prohibited

The Future Of Scalability, Performance, And Reliability

Mike Gualtieri, Senior Analyst, Architecture & Design

February, 2011

Page 3: The future of scaling forrester research - GigaSpaces Road Show 2011
Page 4: The future of scaling forrester research - GigaSpaces Road Show 2011

An explosion of activity and information.

Page 5: The future of scaling forrester research - GigaSpaces Road Show 2011

© 2011 Forrester Research, Inc. Reproduction Prohibited5

Agenda

Massive scale

The seven qualities of wildly desirable applications

Massive scale solutions

Cloud ROI hinges on elasticity

Recommendations

Page 6: The future of scaling forrester research - GigaSpaces Road Show 2011

© 2011 Forrester Research, Inc. Reproduction Prohibited6

Volume equates to transactions, services, and data

19

80

19

81

19

82

19

83

19

84

19

85

19

86

19

87

19

88

19

89

19

90

19

91

19

92

19

93

19

94

19

95

19

96

19

97

19

98

19

99

20

00

20

01

20

02

20

03

20

04

20

05

0

200

400

600

800

1000

1200

1400

1600

1800

Year

Average daily volume of share (million)

Source: NYSE Technologies: Market Data Facts & Figures (http://www.nyxdata.com/Data-Products/Facts-and-Figures)

Average daily volume of shares (in millions) by year on the

New York Stock Exchange (NYSE)

Page 7: The future of scaling forrester research - GigaSpaces Road Show 2011

© 2011 Forrester Research, Inc. Reproduction Prohibited7

More eCommerce means more of everything: users, transactions, and data

Page 8: The future of scaling forrester research - GigaSpaces Road Show 2011

Yikes! This guy can persuade a million

people that my product is rubbish

Page 9: The future of scaling forrester research - GigaSpaces Road Show 2011

© 2011 Forrester Research, Inc. Reproduction Prohibited9

They’re not just kids. Social networking use among Internet users 50 and older has nearly doubled.

Placeholder – social media influence

Page 10: The future of scaling forrester research - GigaSpaces Road Show 2011

© 2011 Forrester Research, Inc. Reproduction Prohibited10

Your business must scale to new heights

• Corporate data warehouses• Historical market data• Structured and unstructured Web data

Huge historical data stores

• Real-time market data• Radio frequency identification (RFID), global position system (GPS)

Exponential increase in business events (data in motion)

• More channels such as eReaders & smart phones (20% of US mobile subscribers now use smart phones)

• Increased expectations for content and transactions

More users

• Business must respond faster• IT must enable continuous business change

Faster pace of change

Page 11: The future of scaling forrester research - GigaSpaces Road Show 2011

© 2011 Forrester Research, Inc. Reproduction Prohibited11

Massive scale examples

Page 12: The future of scaling forrester research - GigaSpaces Road Show 2011

• 145M customer• 20k transactions per second• 1B events daily.

Financial services has an enormous need for scalability, performance, and reliability.

Page 13: The future of scaling forrester research - GigaSpaces Road Show 2011

© 2011 Forrester Research, Inc. Reproduction Prohibited13

Agenda

Massive scale

The seven qualities of wildly desirable software

Massive scale solutions

Cloud ROI hinges on elasticity

Recommendations

Page 14: The future of scaling forrester research - GigaSpaces Road Show 2011

© 2011 Forrester Research, Inc. Reproduction Prohibited14

Wildly desirable business applications must strike the right balance among these seven qualities

Quality What it means

1 ExperienceUsers’ perceptions of the usefulness, usability, and desirability of the application.

2 AvailabilityThe readiness of an application to perform its functions when needed.

3 PerformanceThe speed with which an application performs a function to meet business and user expectations.

4 ScalabilityHandle increasing or decreasing volumes of transactions, services, and data.

5 AdaptabilityThe ease with which an application can be changed or extended.

6 SecuritySupports the security properties of confidentiality, integrity, authentication, authorization, and nonrepudiation.

7 EconomyMinimize the cost to build, operate, and change and application without comprising its business value.

Page 15: The future of scaling forrester research - GigaSpaces Road Show 2011

•Monitor application and infrastructure for expected and unexpected events that can detect or foretell an outage.

Monitored

•Mitigate the effect of known vulnerabilities by implementing redundant components and the ability to operate in a crippled mode.

Fault tolerant

•Identify and rehearse failure scenarios and put in place a support regime to quickly investigate, escalate, and resolve problems.

Fixable

Availability

Page 16: The future of scaling forrester research - GigaSpaces Road Show 2011

© 2011 Forrester Research, Inc. Reproduction Prohibited16

What does high availability mean?

Uptime %* Downtime per year

99.999% (5 nines) 5.26 minutes

99.99% (4 nines) 52.6 minutes

99.5% 1.83 days

99% (2 nines) 3.65 days

98% 7.30 days

95% 18.25 days

*Uptime calculations assume no scheduled downtime.

Page 17: The future of scaling forrester research - GigaSpaces Road Show 2011

© 2011 Forrester Research, Inc. Reproduction Prohibited17

Can “dial tone” reliability be achieved?

Page 18: The future of scaling forrester research - GigaSpaces Road Show 2011

•Application response time meets or beats user expectations and business requirements.

Fast

•Application response time is consistent under normal and load conditions of increasing users, workload, and data.

Consistent

Performance

Page 19: The future of scaling forrester research - GigaSpaces Road Show 2011

© 2011 Forrester Research, Inc. Reproduction Prohibited19

Achieving top performance is a moving target

Source: Gomez.com

Page 20: The future of scaling forrester research - GigaSpaces Road Show 2011

20Entire contents © 2009 Forrester Research, Inc. All rights reserved.

•Application can handle an increasing volume of transactions.

•Example: mouse click to search, button on ATM machine

•Bounding: users, connections, bandwidth

Transactions

•Application can handle an increasing volume of services needed to process transactions.

•Example: business logic, calculations, integration

•Bounding: compute workload

Services

•Applications can handle an increasing volume of transient and persistent data.

•Example: customer profile, transaction history, content, and documents

•Bounding: Size and amount of data

Information

Scalability

Page 21: The future of scaling forrester research - GigaSpaces Road Show 2011

© 2011 Forrester Research, Inc. Reproduction Prohibited21

Agenda

Massive scale

The seven qualities of wildly desirable applications

Massive scale solutions

Cloud ROI hinges on elasticity

Recommendations

Page 22: The future of scaling forrester research - GigaSpaces Road Show 2011

Legacy means you can’t just “rip and replace”.

Page 23: The future of scaling forrester research - GigaSpaces Road Show 2011

© 2011 Forrester Research, Inc. Reproduction Prohibited23

Elastic platforms are a great answer.

Page 24: The future of scaling forrester research - GigaSpaces Road Show 2011

© 2011 Forrester Research, Inc. Reproduction Prohibited24

Elastic application platforms are the future of scalability, performance, and reliability.

Local Cache

Distributed Cache Platform

Elastic Cache Platform

Elastic Application Platform

Memory speeds

Clustered resources

Automatic scaling

Code execution

Fault tolerance

Page 25: The future of scaling forrester research - GigaSpaces Road Show 2011

© 2011 Forrester Research, Inc. Reproduction Prohibited25

Forrester evaluated 8 platforms using 124 criteria

Page 26: The future of scaling forrester research - GigaSpaces Road Show 2011

© 2011 Forrester Research, Inc. Reproduction Prohibited26

Definition: Elastic caching platforms (ECP)

– Platform software that provides data caching

services distributed across two or more server

nodes that:

– consistently perform as volumes grow,

– can be scaled without downtime,

– provide a range of fault-tolerance.

Page 27: The future of scaling forrester research - GigaSpaces Road Show 2011

© 2011 Forrester Research, Inc. Reproduction Prohibited27

Top 6 scenarios for ECPs

Web user session data

Staging of static or transient Web page fragments

Shared application state

Frequently accessed reference data

Result sets from database queries

Compute-intensive jobs on huge data sets

Page 28: The future of scaling forrester research - GigaSpaces Road Show 2011

•Data can be replicated to 2 or more nodes

•Read-through and write-behind can save data to persistent data stores

•Cache can be configured to distribute locks to maintain data integrity during CRUD operations

Availability

•Data stored in-memory for fast access

•Data distributed across 2 or more nodes to avoid bottleneck

•Local cache can store data on same node as code for fastest access

Performance

•Additional nodes can be added in run-time

•Data is re-balanced to take advantage of additional nodes

•Nodes can be removed in real-time to scale down

Scale

Elastic platforms are distributed

Page 29: The future of scaling forrester research - GigaSpaces Road Show 2011

© 2011 Forrester Research, Inc. Reproduction Prohibited29

Typical IT Web architecture statically scales app server tier – but not the data tier at all

Database

Load balancer

Application server

Application server

Application server

Application server

Or, multiple data stores

Page 30: The future of scaling forrester research - GigaSpaces Road Show 2011

© 2011 Forrester Research, Inc. Reproduction Prohibited30

Elastic caching adds a layer to overcome data bottlenecks

Database

Load balancer

Cache node

Cache node

Cache node

Application server

Application server

Application server

Application server

Or, multiple data stores

Page 31: The future of scaling forrester research - GigaSpaces Road Show 2011

© 2011 Forrester Research, Inc. Reproduction Prohibited31

Elastic application adds a layer to overcome both data and application bottlenecks – elastically!

Database

Load balancer

Code & Cache node

Code & Cache node

Code & Cache node

Application server

Application server

Application server

Application server

Or, multiple data stores

Page 32: The future of scaling forrester research - GigaSpaces Road Show 2011

Elastic caching can eliminate bottlenecks.

Page 33: The future of scaling forrester research - GigaSpaces Road Show 2011

© 2011 Forrester Research, Inc. Reproduction Prohibited33

Programming to an elastic cache is straightforward

Scale, performance, & reliability just works

Page 34: The future of scaling forrester research - GigaSpaces Road Show 2011

© 2011 Forrester Research, Inc. Reproduction Prohibited34

Elasticity requires more than just caching

Page 35: The future of scaling forrester research - GigaSpaces Road Show 2011

© 2011 Forrester Research, Inc. Reproduction Prohibited35

Agenda

Massive scale

The seven qualities of wildly desirable applications

Massive scale solutions

Cloud ROI hinges on elasticity

Recommendations

Page 36: The future of scaling forrester research - GigaSpaces Road Show 2011

Cloud computing

Page 37: The future of scaling forrester research - GigaSpaces Road Show 2011

© 2011 Forrester Research, Inc. Reproduction Prohibited37

Steady state of ~40 instances

Launch of Facebook modification

Peak of 5000 instances

The power of cloud economics is elasticity

Page 38: The future of scaling forrester research - GigaSpaces Road Show 2011

© 2011 Forrester Research, Inc. Reproduction Prohibited38

The basics of cloud economics are clear

• As demand rises, resources are added• Requires apps to scale out• Load balancing governs

Elastic scale delivers just-in-time capacity

• No upfront payment• No commitment whether you use it or not• Costs aligned to use

Pay per use keeps costs low

• If you can build it, you can deploy it – fast• API access drives use of automation

Self-service fuels productivity

Page 39: The future of scaling forrester research - GigaSpaces Road Show 2011

© 2011 Forrester Research, Inc. Reproduction Prohibited39

Agenda

Massive scale

The seven qualities of wildly desirable applications

Massive scale solutions

Cloud ROI hinges on elasticity

Recommendations

Page 40: The future of scaling forrester research - GigaSpaces Road Show 2011

Use EAP to achieve high availability.

Replicated, distributed data

Page 41: The future of scaling forrester research - GigaSpaces Road Show 2011

Use EAP to achieve breakneck performance.

In-memory, optimized access

Page 42: The future of scaling forrester research - GigaSpaces Road Show 2011

42Entire contents © 2009 Forrester Research, Inc. All rights reserved.

Use EAP to achieve elastic scale.

Add and remove nodes as needed

Page 43: The future of scaling forrester research - GigaSpaces Road Show 2011

Which of your competitors are doing this now?

Page 44: The future of scaling forrester research - GigaSpaces Road Show 2011

© 2009 Forrester Research, Inc. Reproduction Prohibited

Thank you

Mike [email protected]

Twitter: mgualtieri

www.forrester.com