170 likes | 334 Views
Controlling Data Disclosure in Cross-domain Network Monitoring : Challenges and Approaches. Giuseppe Bianchi Giuseppe.bianchi@uniroma2.it FP7 DEMONS Integrated Project – Scientific Coordinator June 9, 2011 – PoFi Workshop, Pisa. Cross-domain data sharing.
E N D
Controlling Data Disclosure in Cross-domain Network Monitoring:Challenges and Approaches Giuseppe Bianchi Giuseppe.bianchi@uniroma2.it FP7 DEMONS Integrated Project – ScientificCoordinator June 9, 2011 – PoFi Workshop, Pisa
Cross-domain data sharing • Largelyadvocated; supportfor best addressinglarge scale cross-domain threats • Botnet detection and mitigation • DDoS • … • BUT: severe deploymentissues • Business information protection • Customers’ privacyissues • Mismatchesamongtrans-nationalregulation
DEMONS’ Project ChallengesTowards cross-domain cooperation • Disclaimer - technologies alone are notsufficient • Cooperationmeansmustaddresslegalconflicts • Data protectionlaws secureoperationof network services • PlayersmustestablishformalSharingAgreements • Lackof (established/deployed) standards • Cooperationmustbeintensifiedbypoliticalwill and support • Buttechnicalapproachesforsecuring and protectingcooperation and data sharingmayfostercooperation • As enablers, tobuild on top of
DEMONS’ tworesearchdirections(in the secure inter-domain cooperation area) • Tailor secure multiparty computation to network monitoring • SEPIA, P4P, lightweight approaches • Scalable, efficient, suitable for monitoring tasks (mostly additions) • Cooperative Conditionally Encrypted Data Sharing • Share (very selectively) data when anomalous events emerge in multiple domains • Similar alert in multiple domains likely a large scale cross-domain threat permit exchange of relevant data • this talk
Whatwemean(laymanexample) DOMAIN 1 4 out of 5 domains report anomalyinvolvingpxyz.biz pxyz.bizappears (to 4 out of 5 domains) involved in some sortofbotnet fast fluxing… Somethingstrangewithdomainspxyz.biz, base.com Share data relatedtopxyz.biz (data otherwiseconfidential) e.g. full listof end hostsaccessingpxyz.biz, asthesearelikelybotshare among ALL! Also Domain 2 shouldbemadeawareofthis, and Domain2’s pxyz.biz log helpfulhereaswell! DOMAIN 2 Nothingstrangetoday DO NOT share anyremaining log associatedtoother DNS names As thisincludesconfidential/business data, and thereis no evidencefor a global threat or cross-domain issue… Somethingstrangewithdomainspxyz.biz, kristin.it Somethingstrangewith domain pxyz.biz DOMAIN 3 Somethingstrangewithdomainspxyz.biz, alpha.org DOMAIN 4 DOMAIN 5 CONDITIONALLY SHARE onlyif (consistent) anomaliesoccur
Cooperative ConditionallyEncrypted Data Sharing - concept Shared (encrypted) repository DOMAIN 1 Pxyz.biz DOMAIN 2 Alpha.com Beta.it D2 logs Gamma.net … ……… DOMAIN 3 … …… DOMAIN 4 Index Related log Pxyz.biz DOMAIN 5 Alpha.com Beta.it D5 logs Gamma.net … ………
Cooperative ConditionallyEncrypted Data Sharing - concept Pxyz.biz anomaly Shared (encrypted) repository DOMAIN 1 Pxyz.biz DOMAIN 2 Alpha.com Beta.it D2 logs Gamma.net … ……… Pxyz.biz anomaly DOMAIN 3 … …… Pxyz.biz anomaly DOMAIN 4 Pxyz.biz DOMAIN 5 Alpha.com Pxyz.biz anomaly Beta.it D5 logs Gamma.net … ………
Cooperative ConditionallyEncrypted Data Sharing - concept Shared (encrypted) repository DOMAIN 1 Pxyz.biz DOMAIN 2 Alpha.com Beta.it D2 logs Gamma.net … ……… DOMAIN 3 … …… DOMAIN 4 Pxyz.biz DOMAIN 5 Alpha.com Beta.it D5 logs Gamma.net … ………
Technicalhassles • LOTS ofpossibleindexes/logs per name, per flow, … millions?! forgetexplicitper-flowcooperation • Symmetricencryption performance! Tocopewith potentially massive logs and potentiallyhugenumber • Distinctencryptionkeys game over, otherwise (allmightread) mustautomate key management… • No trustedthird party control unviable, not a real world stakeholder • Lookslike secret sharing, but.. whoprovides the secret? one secret? or one per index?! from secret(s) toindependentper-domain & per-indexkeys: howto?
Proposedconstruction (at a glance) • Asymmetricencryptionofsymmetrickeys • Useordinarysymmetriccipherforlogs (e.g. AES-128) • Delivereachper-logAES-keycoveredbyasymmetricencryption • Problembecomes: howtoescrowsuchencryptedAES-keys • Pedersen’s Distributed Key Sharing • Setup “unknown” shared secret S amongdomains (verifiable SS) • Gettoknowrelated public valuegS • Boneh & Franklin’s Pairing-basedIdentityBasedEncryption • Use public valuegSas public IBE key (no PKG needed, obviously) • Encryptper-identity, whereidentity ID is the indexofeach log • Escrow private (per-identity) key byreleasingsharesof IDS
Details (1/5): Setup • Once-for-all Domain cooperation via Pedersen DKG • Secret S generated • uniquebutunknowntoall • Domain i onlyknowsone share xiofit • gSbecomes public • gS = Pedersen’sverificationvaluewithindex 0 • Domainsagree on non degenerate computablebilinearpairing: GG G1 finite cyclic groups order q s.t.:e(ga, gb) = e(g,g)aba,bZ, gG
Details (2/5): per-indexencryption • Index = arbitraryfiltering key • DNS name, IP address, Flow tuple, etc • Log = information associatedtoindex • Listofqueryinghosts, deliveredpackets, etc • Cipher = symmetric = e.g., AES-128 • Encryptall log associatedtoindex • Automatedpseudorandom key generation (stateless)onedistinct key per index ID and per domain i: Ki,ID=K(domain i, index ID)=PRF(Domain secret, ID)
Details (3/5): publish IBE encryptedper-index key • Indexidentity: IDf=Hash(indexname) over G • Release in public (once per log): • gr r (pseudo)random • Again, PRF tricktomakeitstateless • Encrypted key = Ki,f e(gs, (IDf)r)= Ki,f e(g,IDf)rs • Usual public key encryption: private key neededfordecrypt • butnone hasit, at this stage
Details (4/5): release key share upondetectedanomaly • Byconstruction, domain has a share xiof secret S • For the “anomalous” index, domain i derives a per-index share: (IDf)xi • RememberB&F IBE: (IDf)S is the private key forIDf • (IDf)xi = (exponent) share of private IBE key • Whentorelease share: localanomaly detection mechanism (any appropriate to the scope at hands) • Example: DNS analysis fast-flux DNS candidate
Details (5/5): Escrowkeys • Anyone (in principle) in sharedrepositorymaycollectshares • Whensufficientnumberofsharesavailable, reconstruct IBE private key forIDf via (exponent) Lagrangeinterpolation • Lagrange coeff. dependingonlyupon domain indexes (known) • Decryptusingusualpairing: e(gr, IDfs)=e(g, IDf)rs
Discussion • Security: founded on B&F IBE provable security under private key disclosure • Knowing private keysfor a set ofIDsdoesnotpermittodecryptanyother ID • Performance: slow pairingdone once per index • Massive log encryption via fast ordinary AES-128 symmetric • Scalability: automated (stateless) key generation, no maintenance • Overhead@setup • Notcritical, once-for-all • Allremainingoperationdoes NOT requireanycoordination • Rekeying: involvesonlysymmetrickeys, no DKG coordination
Issues & nextsteps • Currently just “pure” threshold-based (t,n) approach • ongoing work: support more expressiveaccesscontrol (data sharing) policies • Decrypt on the basisofmore specificalerts (or differenttypesof) fromselectedsubsetsofdomains • generalizeto monotone accessstructures via linear secret sharingschemes • mutuate ideas/approachesfrom (decentralized) CP-ABE • Proof-of-conceptreal-worldmonitoringuse-case: long term target • Current candidate scenario: cooperative botnet detection; • Needtodevelop (complementary) DNS analysisforalert generation