MODAClouds Decision Support System for Cloud Service Selection
-
Upload
ldbc-council -
Category
Technology
-
view
49 -
download
0
Transcript of MODAClouds Decision Support System for Cloud Service Selection
MODAClouds Decision Support System for Cloud Service Selec8on Smra8 Gupta CA Labs, CA Technologies
20th of March 2015 LDBC Sixth TUC Mee8ng, UPC, Barcelona
2 © 2015 CA. ALL RIGHTS RESERVED.
Outline
Objec8ve of the talk
Need for Decision Support System in Cloud service selec8on
Overview of MODAClouds DSS
Key Features of DSS
Open Discussions for DSS in graph database community
3 © 2015 CA. ALL RIGHTS RESERVED.
Why are we here? Decision Support System and graph databases
CALabs Barcelona team has organically developed a novel technology in the form of Decision Support System as a part of MODAClouds project.
Graph database community is evolving and there lies poten8al to use the DSS technology in addressing the graph database selec8on problem
Objec8ve of this talk is to start brainstorming in the community about possible usage of the technology to assist and enhance the use of graph databases in enterprises
4 © 2015 CA. ALL RIGHTS RESERVED.
Need for Decision Support System in cloud service selec8on
Mul8ple dimensions of choices • Trustworthy Vendors • Financial, Legal, Organiza8onal and Technical constraints
Mul8-‐cloud environment compa8bility issues • Interoperability • Ease of migra8on • Vendor lock-‐in
Recommenda8on based on different dimensions • Cost • Quality • Risk
5 © 2015 CA. ALL RIGHTS RESERVED.
What DSS does for the users?
MODA Clouds DSS
Architectural model of deployment (Tangible Assets)
Architectural deployment model enriched with user selected cloud services
MODAClouds User
Cloud Service Recommenda8ons
Technical and Business oriented Intangible assets and Risk Acceptability level per asset
Relevant Risks and Treatments
Selected cloud service alterna8ves
6 © 2015 CA. ALL RIGHTS RESERVED.
MODAClouds DSS: Key features
§ Mul8ple Stakeholder par8cipa8on
§ Risk-‐analysis based Requirement genera8on
§ Mul8-‐Cloud Environment Compa8bility
§ Data gathering § Progressive Learning
7 © 2015 CA. ALL RIGHTS RESERVED.
Mul8ple actors, mul8ple perspec8ves
§ Different stakeholders may influence Cloud Service selec8on in different ways
Risk Policy Manager
Decision Owner
Architect
System Operator
Feasibility Study
Engineer
7
8 © 2015 CA. ALL RIGHTS RESERVED.
Asset defini8on by mul8ple actors
Business Analyst
Assets
Product Innova8on and Quality
Legisla8on Compliance
Sales Rate Customer Loyalty
Market Awareness
Business-Oriented Intangible Assets
8
Technical-Oriented Intangible Assets
Assets
Data Privacy
Data Integrity
End User Performance
Maintainability
Service Availability
Cost stability
Technical Team
Assets
Compute (IaaS)
File System (IaaS)
Blob storage (IaaS)
Rela8onal (PaaS)
Middleware (PaaS)
NoSQL (PaaS)
Backend (PaaS)
Frontend (PaaS)
9 © 2015 CA. ALL RIGHTS RESERVED.
Risk analysis methodology
Business Oriented Intangible Asset
Defini8on
Technical Oriented
Intangible Asset Defini8on
Tangible Assets Defini8on Risk defini8on Treatments
Defini8on
§ Risks are iden8fied on the basis of protec8ng the assets § Treatments are defined to mi8gate one or more risks
§ The outputs can be refined itera8vely allowing users to go back in the methodology and update informa8on
9
10 © 2015 CA. ALL RIGHTS RESERVED.
Mul8-‐Cloud environment
11 © 2015 CA. ALL RIGHTS RESERVED.
Challenges in Mul8-‐Clouds
11
• Interoperability: Risk of unexpected lack of replacement and consequent vendor lock-‐in • Migra8on: Risk of non-‐viable migra8on due to migra8on costs and complexity Vendor lock-‐in
• Risk of new security breaches due to the increased complexity of the system and new communica8ons Security
• Risk of unavailability of evidences in case of fraudulent ac8ons Forensic Evidences
• Risk of costs unpredictability Cost unpredictability
• Risk of lack of provider interest in collabora8on Lack of interest of CSPs
• SME or companies using mul8ple services from mul8ple vendors are unlikely to have the power or the 8me to nego8ate. Increasingly unstable cost and T&C problem.
Lack of nego8a8on on SLAs capacity
12 © 2015 CA. ALL RIGHTS RESERVED.
DSS – Automa8c Data Gathering Concept
DSS Database
Graph building and data
transforma8on
Structured flat data fetch
JSON Database Interface
XML
REST
JSON
XLSX
WSDL
NoSQL SQL
Internet
Flat files
Databases
Graph
13 © 2015 CA. ALL RIGHTS RESERVED.
Progressive Learning
Storage of User input
Storage of selec8on of services
Storage of thresholds
and benchmarks
Subsequent recommend-‐a8on on selec8on
Subsequent recommenda8on on services
• With repeated use of DSS, the previous user logs and stored and simple analysis is performed
• The recurring users are recommended possible assets that might be crucial to their firm
• The users are also recommended certain risks that have been chosen by other users
• The users are also recommended the value of each cloud service property based on previous use of DSS
• With the repeated usage, DSS learns and improves its recommenda8ons
14 © 2015 CA. ALL RIGHTS RESERVED.
Ground-‐up developed Prototype by CALabs
15 © 2015 CA. ALL RIGHTS RESERVED.
Open Source Technology Support for DSS
• hmp://dss.tools.modaclouds.eu/ DSS open source tool available at:
• hmps://github.com/CA-‐Labs/DSS Documented and available in github repository at:
• hmp://www.modaclouds.eu/ MODAClouds Documenta8on
16 © 2015 CA. ALL RIGHTS RESERVED.
Open Discussion -‐ What are the characteris8cs that would define the quality of a cloud graph database? -‐ What criteria are important in the selec8on of (cloud) graph databases?
Who makes the decisions in industry to select a par8cular graph database technology for a company?
How does the graph database community plan to manage legi8mate customer concerns such as preven8on of vendor lock-‐in and cloud outages? Is the synchroniza8on of mul8ple graph databases provided by different vendors possible?
Is gathering data with respect to different characteris8cs that define the quality of the graph database an important concern?
How could a DSS help for cloud graph database selec8on?
17 © 2015 CA. ALL RIGHTS RESERVED.
Thank you for your amen8on!
Sr. Research Engineer [email protected]
Dr. Smra8 Gupta