epts survey results

24
epts event processing technical society epts event processing technical society Pedro Bizarro on behalf of the Use Case Working Group Results of the Survey on Event Processing Use Cases March 24, 2011

Upload: pedro-bizarro

Post on 29-Nov-2014

963 views

Category:

Documents


0 download

DESCRIPTION

Presentation of the Event Processing Survey prepared by the Use Cases Workgroup of the Event Processing Technical Society. Presented at the 6th EPTS Symposium at March 24, 2011

TRANSCRIPT

Page 1: EPTS Survey results

epts event processing technical society

eptsevent processing technical society

Pedro Bizarro on behalf of theUse Case Working Group

Results of the Survey on Event Processing Use CasesMarch 24, 2011

Page 2: EPTS Survey results

epts event processing technical society

2

How are event processing technologiesbeing used?

Classify scenarios,help othersselect solutions

Inspire use ofevent processing

Food for thought forresearchers & engineers

Page 3: EPTS Survey results

epts event processing technical society

3

Earlier versions

• 2007: Kick-off – Problem: world of superficial use cases

• 2008: v1: 6 use cases, 54-questions (@4th epts)

• 2009: v2: 5 use cases 9 lessons learned (@debs2009)

• 2011: v3: 30 use cases plenty of statistics (today!)

Page 4: EPTS Survey results

epts event processing technical society

4

24 questions~13 minutes

30 use casesEPTS, DEBS

community

Page 5: EPTS Survey results

epts event processing technical society

Large Variety of Use Cases

5

GridDocument workflow

Patient discharge

Revenues and expensesETL in Telcos

NYC transportation

Content authoring

Gas station networks

Emergency management

Home energy

Testing algorithms…

Page 6: EPTS Survey results

epts event processing technical society

6

Banking/Financial

Energy/Utilities

Aerospace/Defense

Education

Telecom

Transportation/Logistics

Other

0% 5% 10% 15% 20% 25%25%

18%

11%

7%

7%

7%

25%

Industry background

Computer SoftwareHealthcareManufacturingRetail and distributionetc<4% each

Page 7: EPTS Survey results

epts event processing technical society

7

Functional area

Operations

Research and Development

Finance

IT Support / Administration

Customer Service and Support

Human Resources

Sales

Marketing

Other

0% 5% 10% 15% 20% 25% 30% 35%

33%

18%

12%

12%

9%

6%

3%

0%

6%

Page 8: EPTS Survey results

epts event processing technical society

8

Maturity level

A concept

Prototyping - any stage

Pilot

In developement

In production

0% 5% 10% 15% 20% 25% 30% 35%

4%

17%

26%

22%

30%

52%

Page 9: EPTS Survey results

epts event processing technical society

9

Primary project drivers

Enhance/improve user services

Add new features

Reduce production cost

Faster user service

Reduce deployment cost

Compliance

Other

0% 5% 10% 15% 20% 25% 30% 35% 40%

35%

16%

14%

11%

8%

5%

11%

Page 10: EPTS Survey results

epts event processing technical society

10

Data sources

Databases

Files

ESB, JMS, other msg protocols

RSS, twitter, stock, news, other feeds

Application or OS events

Video, sound, multimedia streams

Other

0% 5% 10% 15% 20% 25% 30% 35%

31%

18%

18%

10%

5%

0%

18%

Non“streaming”

“streaming”

Page 11: EPTS Survey results

epts event processing technical society

11

Destinations and actions

Msg people (E-Mail, SMS, MMS)

Trigger other applications

Databases

Dashboards

ESB, JMS, other msg protocols.

RSS, twitter, stock, news, other feeds

Other

0% 5% 10% 15% 20% 25% 30%

24%

24%

20%

17%

7%

4%

4%

someautomation

but peoplestill stronglyin the loop

Page 12: EPTS Survey results

epts event processing technical society

12

Desired features

Correlations or joins

Summaries or aggregations

Sliding windows

Tumbling windows

Forecasting and predictive functions

Probability or statistical models

Other

0% 5% 10% 15% 20% 25% 30%

24%

24%

17%

16%

9%

9%

2%

simplerbeats

complex

Page 13: EPTS Survey results

epts event processing technical society

13

Data models and data types

Relational

Column-Oriented

Generic XML

Objects

RDF

XML (industry specific schema)

Extensibility

Other

0% 5% 10% 15% 20% 25% 30% 35%

32%

18%

12%

9%

9%

9%

3%

9%

Page 14: EPTS Survey results

epts event processing technical society

14

Performance – input events per second

NONE

'1-10

'10-100

'100-1.000

'10.000-100.000

'100.000-1.000.000

1.000.000 or more

No answer

0% 5% 10% 15% 20% 25% 30% 35% 40%

5%

11%

16%

37%

11%

5%

0%

16%

Not that much!

Page 15: EPTS Survey results

epts event processing technical society

15

Number of data sources

'1-10

'10-100

'100-1.000

'10.000-100.000

'100.000-1.000.000

1.000.000 or more

No answer

0% 5% 10% 15% 20% 25% 30% 35% 40% 45%

42%

26%

11%

0%

5%

0%

16%

Or maybe not! Typo?!

Data comes fromfew sources

Page 16: EPTS Survey results

epts event processing technical society

16

Number of “AI” models

NONE

'1-10

'10-100

'100-1.000

'10.000-100.000

'100.000-1.000.000

1.000.000 or more

No answer

0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50%

47%

16%

11%

0%

0%

0%

0%

26%

3 in 4 don’t answeror don’t use

Page 17: EPTS Survey results

epts event processing technical society

17

NONE

'1-10

'10-100

'100-1.000

'1.000-10.000

'10.000-more

No answer

0% 5% 10% 15% 20% 25% 30% 35% 40%

5%

11%

16%

11%

5%

16%

37%

Expected data size growth rate/year (in MB)

or not growing much

Hard to forecast

47% increase lessthan 10Gb/year

Page 18: EPTS Survey results

epts event processing technical society

18

Why did we use log scales?

Because we didn’t knowwhat to expect

Page 19: EPTS Survey results

epts event processing technical society

19

Enterprise capabilities

High availability

Online system maintenance

Provenance

Fault tolerance

Recovery of any data or meta-data

Security

Encryption

0% 5% 10% 15% 20% 25% 30% 35% 40% 45%

39%

24%

24%

21%

21%

12%

0%

Cannot stop!

Not a big concern

Page 20: EPTS Survey results

epts event processing technical society

20

Implementation constraints

Specific CEP engine

Specific file or database system

Specific operating system

Specific application server

Specific development environment

Specific hardware

Other

0% 5% 10% 15% 20% 25% 30%

27%

19%

15%

12%

12%

8%

8%

surprise

Page 21: EPTS Survey results

epts event processing technical society

21

The typical use case

• In production to improve banking/utilities user services

• Gets data from databases, files and message queues

• Notifies people and other applications

• Does correlations, joins and aggregations

• Handles less than 1000 events/second (<10 sources)

• Must run a specific CEP engine

• Cannot stop

Page 22: EPTS Survey results

epts event processing technical society

22

The surprises (personal take)

• Few telcos, transportation and logistics

• Not from IT department

• Not about lower TCO, faster deployments

• “Non-streaming” dominate

• Low throughput, few sources, little growth

• Not about forecasting, predictions

• Very few AI models

• Not about security, encryption

Page 23: EPTS Survey results

epts event processing technical society

23

Members and acknowledgements

• Pedro Bizarro <[email protected]>

• Christoph Emmersberger <[email protected]>

• Thomas Ertlmaier <[email protected]>

• Matthew Cooper <[email protected]>

• Tina Groves <[email protected]>

• Dieter Gawlick <[email protected]>

• Brian Connell <[email protected]>

Page 24: EPTS Survey results

epts event processing technical society

eptsevent processing technical society

Q&A?

24