Principles of Elastic Processes on Clouds and Some Enabling Techniques

ASIS' 11, Wollongong, Nov 22, 2011 1

Principles of Elastic Processes on Clouds and Some Enabling Techniques

Hong-Linh Truong

Distributed Systems Group, Vienna University of Technology

[email protected]://www.infosys.tuwien.ac.at/Staff/truong

Joint work with: Schahram Dustdar, Frank Leymann, Michael Reiter, Tran Vu Pham, Quang Hieu Vu, Yike Guo, Benjamin Satzger, Uwe Breitenbuecher, Dimka Karastoyanova

Presented at Australian Symposium on Services InnovationNovember 21st and 22nd, 2011, University of Wollongong

mailto:[email protected]

ASSI' 11, Wollongong, Nov 22, 2011 2

Outline

Principles of elastic processes

Some enabling techniques for elastic processes in hybrid systems Monitoring and analysis of non-functional parameters


Data-as-a-service and computation-as-a-service

Multiple realtime data sources – data services Smart environments Earth monitoring, satellite images Social media Web information Personal medical information

Data processing algorithms – computing services Data conversion, data enrichment, data extraction, data

mining, domain-specific computational analysis (e.g., CO2 calculation), etc

Algorithms can be wrapped into processing elements/workflow activities


Data intensive processes Influence factors for the quality of processed

results Data rates, quality of data sources, underlying

computing systems, etc.

The quality of the output is varying Multiple concurrent consumers

Dynamic requirements for SLA, quality of outputs, etc.

Data processing service provider: my process must be elastic! to optimize resource usage, prices, etc., but must meet

consumer's requirements


Example of a business/e-science process

algo1

Outputs: results,

accuracy,performance

etc.

algo2

algo3

algo4 algo5

=?Reqs:

Consumer 1,Consumer 2Consumer n

Resources in multiple clouds

Resources, quality and cost?

Influence factors:

performancequality of data

Quality/cost influence?

algo5

algo6

algo8algo5

algo7

algo4

Quality of data?

DaaS


Elasticity in computing

„Elastic computing is the use of computer resources which vary dynamically to meet a variable workload” – http://en.wikipedia.org/wiki/Elastic_computing

„What elasticity means to cloud users is that they should design their applications to scale their resource requirements up and down whenever possible.“, David Chiu – http://xrds.acm.org/article.cfm?aid=1734162

„Clustering elasticity is the ease of adding or removing nodes from the distributed data store” – http://en.wikipedia.org/wiki/Elasticity_(data_store)

http://en.wikipedia.org/wiki/Elastic_computing


Elasticity in computing – a broad view in multi-cloud environments

Elastic demands from consumers Multiple outputs with different price and quality (output

elasticity) Elastic data inputs, e.g., deal with opportunistic data Elastic pricing and quality models associated resources

The service/process itself is elastic in terms of economic theory

The service/process structure is elastic in terms of physics


Multiple elasticity dimensions

Resource elasticity: software/human-based computing elements, multiple clouds

Non-functional parameters: performance, quality of data, service availability, human trust, ...

Pricing/Rewarding/Incentive elasticity

Elasticity


Conceptual architecture of elastic process environment

Schahram Dustdar, Yike Guo, Benjamin Satzger, Hong Linh Truong: Principles of Elastic Processes. IEEE Internet Computing 15(5): 66-71 (2011)


Elastic process – operation principles

Operation principles Monitor, manage and describe dynamic properties Dynamically refine process functions based on

quality Determine costs based on multiple resource

price/reward/incentive models Provide elasticity across multiple cloud providers

An EP can deal with multiple service objectives N concurrent consumers, access markets of M

providers, with K different requirements: K <= N


Elastic process – modeling principles

Overlaying EPs, functional composition, and dynamic property composition

Modeling principles EP's function as a static property EP's results based on requirements concerning

price/reward/incentive and quality modeled as a set of constraints influencing resource

elasticity

Communication can be based on SOAP/REST Refinement and composition of resource, price and

quality at multiple levels: activities, fragments, or the whole EP


The Vienna Elastic Computing ModelE-science Application Business Application

Elastic Service

Elastic Alignment

Elastic Runtime Management

IT Elastic Service

IT Elastic Process

Software-based Computing Element

(SE)

Virtualization

Human-based Computing Element

(HE)

Heavy crowded in cloud computing and social computing

Ready to use but in isolation (single cloud and single type of computing elements)

Service modeling and interface

Alignment techniques

Virtualization techniques

Monitoring and analysis techniques

Non-functional parameters characterization

Programming models

Governance and service evolution


Clouds of hybrid systems

Software- and human-basedresource virtualization

Middleware

Modeling and alignment

Evolution

Programming languages

Refinement and adaptation

NFP Characterization, Monitoring and Analysis

Pricing/Rewarding/Incentive mechanisms

Governance and compliance


How to determine the cost model of a complex process?

Coarse- and fine-grainted cost models of clouds at different layers:

Too coarse-grained (networks, storages, machines) or too fine-grained (IO calls)

Application-, software-, data-, human-specific cost/performance models

Cost models for individual parts (workflow, MPI, OpenMP, etc.)

Elastic cost models associated with resources and services in multiple clouds

Part A

Part B

Part C

Elastic Process

cost/performance model i

cost/performance model j

cost/performance model k

Our approach


How to monitor software based computing elements across clouds?

Hong Linh Truong, Schahram Dustdar: Composable cost estimation and monitoring for computational applications in cloud computing environments. Procedia CS 1(1): 2175-2184 (2010)

Tran Vu Pham, Hong-Linh Truong, Schahram Dustdar "Elastic High Performance Applications - A Composition Framework", The 2011 Asia-Pacific Services Computing Conference (IEEE APSCC 2011), (c) IEEE Computer Society, December 12 - 15, 2011, Jeju, Korea

Elastic high performance applications on multiple clouds


Are current “software service techniques” suitable for DaaS/data marketplaces

Data marketplaces and DaaSElastic dynamic properties associated with:Service APIs, data APIs and data assets

- Quang Hieu Vu, Tran-Vu Pham, Hong-Linh Truong, Schahram Dustdar, Rasool Asal, DEMODS: A Description Model for Data-as-a-Service. AINA 2012. To appear.- Hong-Linh Truong, Schahram Dustdar "On Analyzing and Specifying Concerns for Data as a Service", The 2009 Asia-Pacific Services Computing Conference (IEEE APSCC 2009), (c) IEEE Computer Society, December 7-11, 2009, Biopolis, Singapore.- Marco Comerio, Hong-Linh Truong, Flavio De Paoli, Schahram Dustdar, " Evaluating Contract Compatibility for Service Composition in The SeCO2 Framework ", The 9th International Conference on Service Oriented Computing (ICSOC 2009), (c) Springer-Verlag, November 24 - 27, 2009, Stockholm, Sweden.


How to monitor quality of data in an elastic way?

Pull, pass-by-references Pull, pass-by-values

Push, pass-by-values

Hong Linh Truong, Schahram Dustdar: On Evaluating and Publishing Data Concerns for Data as a Service. APSCC 2010: 363-370


How to deal with subjective and objective evaluation in long running processes?

Our approach Different QoD

measurements Human and software

tasks


Experiments: cross-layered monitoring in multiple clouds

Cloud provider bill

Online analysis Estimation


Experiments: evaluating and publishing QoD with data resources


Experiments: evaluating quality of data in workflows

Michael Reiter, Uwe Breitenbuecher, Schahram Dustdar, Dimka Karastoyanova, Frank Leymann, Hong-Linh Truong, A Novel Framework for Monitoring and Analyzing Quality of Data in Simulation Workflows, (c)IEEE Computer Society, The 7th IEEE International Conference on e-Science, 5-8 December, 2011, Stockholm, Sweden


Open questions for enabling techniques in hybrid systems

Software, data and human integration on multiple clouds

Monitor elastic processes in hybrid systems Provisioning on-the-fly monitoring tools for complex cost models from

different layers

Monitoring human-based computing elements

Check contract compatibility and compliance – no static Evaluating dynamic contract of data, software-based services, and

human-based computing elements in the same process

Evaluate quality of data in complex workflows Modeling hybrid processes of human and software for evaluating quality

of data in data intensive workflows

Utilizing quality of data in supporting the elasticity of cloud service and infrastructure provisioning


Summary

Elastic computing is needed in different aspects Scaling software, services, and people in the same

application

Not just “resource elasticity” Resource, price/reward/incentive and quality Hybrid systems of software-based and human-based

computing elements Multiple clouds

Several open research questions for realizing elastic processes of hybrid computing elements


Thanks for your attention!

Hong-Linh TruongDistributed Systems GroupVienna University of TechnologyAustria

[email protected]://www.infosys.tuwien.ac.at

mailto:[email protected]

Documents

Principles of Elastic Processes on Clouds and Some Enabling Techniques