12
Grid Computing Grid Computing Security Security A Taxonomy A Taxonomy Fletcher Liverance, 5 May 2009 IEEE Security & Privacy, 2007 Anirban Chakrabarti Anish Damodaran Shubhashis Sengupta

Grid Computing Security

  • Upload
    wright

  • View
    33

  • Download
    0

Embed Size (px)

DESCRIPTION

Fletcher Liverance, 5 May 2009. Grid Computing Security. A Taxonomy. IEEE Security & Privacy, 2007 Anirban Chakrabarti Anish Damodaran Shubhashis Sengupta. Overview. What is Grid Computing? Pie in the sky Host-level issues and solutions Architecture-level issues and solutions - PowerPoint PPT Presentation

Citation preview

Page 1: Grid Computing Security

Grid Computing SecurityGrid Computing Security

A TaxonomyA Taxonomy

Fletcher Liverance, 5 May 2009

IEEE Security & Privacy, 2007 Anirban ChakrabartiAnish DamodaranShubhashis Sengupta

Page 2: Grid Computing Security

OverviewOverview

What is Grid Computing?What is Grid Computing?

Pie in the skyPie in the sky

Host-level issues and solutionsHost-level issues and solutions

Architecture-level issues and solutionsArchitecture-level issues and solutions

Credential-level issuesCredential-level issues

Page 3: Grid Computing Security

What is Grid Computing?What is Grid Computing?

““geographically distributed heterogeneous geographically distributed heterogeneous resources are virtualized as a unified whole.”resources are virtualized as a unified whole.”

• Web 2.0• Scalable Link Interface (SLI)• Virtualization• Software as a service• Folding@home• Peer to peer• Cluster computing• Cloud computing• Distributed computing

Page 4: Grid Computing Security

Computing ComparisonComputing Comparison

Page 5: Grid Computing Security

Pie in the skyPie in the sky

IBM RoadrunnerIBM Roadrunner 6,480 AMD dual core6,480 AMD dual core 12,960 IBM PowerXCell12,960 IBM PowerXCell

Hewlett-PackardHewlett-Packard 300,000 employees300,000 employees 600,000 processors600,000 processors 600 TB of RAM600 TB of RAM 120,000 TB of Storage120,000 TB of Storage

World wideWorld wide One billion PCsOne billion PCs 95 million consoles95 million consoles Two billion cell phonesTwo billion cell phones

Page 6: Grid Computing Security

Host-level issues and solutionsHost-level issues and solutions

Data ProtectionData Protection Application-level sandboxingApplication-level sandboxing

Proof-carrying codeProof-carrying code Rules guaranteeing safe executionRules guaranteeing safe execution Code producer responsible for safetyCode producer responsible for safety Does not scaleDoes not scale

VirtualizationVirtualizationVMware GSX/ESX/WorkstationVMware GSX/ESX/WorkstationParavirtualizationParavirtualization

XenXen IA-32 architecture is non-virtualizableIA-32 architecture is non-virtualizable

Page 7: Grid Computing Security

Host-level issues and solutionsHost-level issues and solutions

Data Protection Data Protection User-space sandboxingUser-space sandboxing

TRON – Process-level discretionary access control systemTRON – Process-level discretionary access control system

Simple, but requires system call reimplementationSimple, but requires system call reimplementation

Call chaining issuesCall chaining issues

Incomplete contextIncomplete context Flexible kernels (Kernel-level sandboxing)Flexible kernels (Kernel-level sandboxing)

Exokernel OS, MITExokernel OS, MIT

Zones, Sun Solaris 10Zones, Sun Solaris 10

Application containersApplication containers

Page 8: Grid Computing Security

Host-level issues and solutionsHost-level issues and solutions

Job starvationJob starvation Advanced reservation techniquesAdvanced reservation techniques

Request resources from grid schedulerRequest resources from grid scheduler

Non-transparentNon-transparent

Requires advanced scheduling techniquesRequires advanced scheduling techniques Priority-reduction techniquesPriority-reduction techniques

Local priority reductionLocal priority reduction

Sun Grid engineSun Grid engine

Ad hoc mechanismAd hoc mechanism

Unpredictable behaviour, lower QoS performanceUnpredictable behaviour, lower QoS performance

Example: Peer to peerExample: Peer to peer

Page 9: Grid Computing Security

Architecture-level issues and Architecture-level issues and solutionssolutions

Information securityInformation security Grid Security Infrastructure (GSI)Grid Security Infrastructure (GSI) Secure communicationSecure communication

Transport level security - SSL/TLSTransport level security - SSL/TLSMessage level security – Web Services Security (WSS) via Message level security – Web Services Security (WSS) via SOAPSOAP

AuthenticationAuthenticationCA CertificatesCA CertificatesUser/password over SOAP with WSSUser/password over SOAP with WSSGSI-to-Kerberos gatewayGSI-to-Kerberos gateway

Single sign-on and delegationSingle sign-on and delegationTimed proxyTimed proxy

Page 10: Grid Computing Security

Architecture-level issues and Architecture-level issues and solutionssolutions

Policy-mapping issuesPolicy-mapping issues Resource levelResource level

Akenti – Distributed access control mechanismAkenti – Distributed access control mechanism Use-condition certificatesUse-condition certificates Attribute certificatesAttribute certificates

Virtual Organization levelVirtual Organization levelCommunity Authorization Service (CAS)Community Authorization Service (CAS)Role based access controlRole based access control

DoSDoS Preventative solutionsPreventative solutions

Application filteringApplication filteringSnort - Intrusion Detection SystemSnort - Intrusion Detection System

Reactive solutionsReactive solutionsLink testingLink testingLoggingLogging

Page 11: Grid Computing Security

Credential-level issuesCredential-level issues

Credential repositoriesCredential repositories Take responsibility for credential storageTake responsibility for credential storage MyProxy OnlineMyProxy Online

Credential federation systemsCredential federation systems ““Manage credentials across multiple systems, Manage credentials across multiple systems,

domains, and realms.”domains, and realms.” KX.509KX.509 Circle of trustCircle of trust ShibbolethShibboleth

Page 12: Grid Computing Security

ConclusionsConclusions

““Grid security’s ultimate goal is to make the Grid security’s ultimate goal is to make the grid infrastructure seamless and protect it grid infrastructure seamless and protect it against both known and unknown security against both known and unknown security attacks.”attacks.”

1.1. Identify vulnerabilitiesIdentify vulnerabilities2.2. Develop threat modelsDevelop threat models3.3. Develop countermeasures to threat modelsDevelop countermeasures to threat models4.4. Evaluate counter measuresEvaluate counter measures5.5. (repeat ad nauseam)(repeat ad nauseam)