Conflict Resolution & Policy Compliance in Multi-Cloud Distributed System. Presented By:- Adarsh Pillay Deepak Begrajka Rudra gupta
Content Introduction to Cloud Computing. Multi Cloud. Threats. Scenarios. Proposed Solution. Conclusion.
Evolution of cloud Computing 1950s Mainframe Computing 1970s Concept of Virtual Machine 1990s Telecommunications companies started offering VPN connections Grid Computing Utility Computing SaaS Cloud Computing
What Is Cloud Computing
Services of Cloud
Limitations in Single cloud Downtime Security and Privacy Vulnerability to attacks Limited control and Flexibility Limited number of Resources Platform Dependencies Cloud Computing Cost
Single Cloud to Multi Cloud
What is multi cloud Use of multiple cloud computing services in a single heterogeneous architecture.
Need of multi cloud Autonomy-: Reducing Reliance on any single vendor Increasing flexibility through choice Mitigation against disaster Hybridity-: Some applications on-premises and others on one or more public clouds Extended capabilities Leveraging public cloud cost benefits Support varying level of security concerns
Threats in multi cloud Increase in attack surface due to system complexity. Data privacy concerns due to multitenancy. Cloud service provider don’t understand each other. Conflict in policies. Loss of client’s control over resources and data due to asset migration. The biggest issue in multi cloud environment is trust. Complexity. Management overhead.
SCENARIO 1 Two or more cloud merge together to form a multi cloud.
SCENARIO 2 New cloud want to join multi cloud.
Trust
incentive
Proposed Solution Weighted Probability Trust. Customization as a Service. Third Party Multi Cloud Audit
Weighted probability Trust Weighted probability= Pi. Number of Interaction for service = M. X, Y are User & A, B, C are CSP. Trust = (USERx, Number of interaction). Trust Factor = Σ(in) M(i)*Pi / Σ(in) M(i).
Example Trust factor of A = Mx * Px + My * Py + Mb * Pb Mx + My + Mb .5 .1 Trust factor of A= 1*.1 + 1*.5 + 1*(-.2) 1+1+1 = .134 -.2
Trusted Certified User A customer giving multiple reviews about multiple users. Ex:- Just like Amazon Certified Buyer.
Weighted Trust Probability + Recommender System CSP 1 want to collaborate with particular service of cloud 2. CSP 1 use recommender system to find trusted cloud. Sign a SLA if both agree. Proceed forward.
Customization Policy as A Service CSPs can choose their own policy and agree on it. User can select any predefined set of policies based on their requirements. Old policy can be used or rebuild new ones according to the requirements and if satisfied incentive.
Trusted third party Multi cloud audit Weighted probability trust has deployed a trusted environment. SLA is there but still we need a physical entity to keep regular check. Third party cloud auditor is the trusted party or friend who will keep regular check in the multi cloud environment.
Conclusion Started with single cloud, But not able to satisfy users need. Multi Cloud is able to satisfy user needs, but still poses some threats and limitations. Trust is the key issue in multi cloud architecture. We propose three method to generate full of trust architecture. Our architecture and methods if implemented can provide high level secure and conflict environment.