Hello from CS - JAIST...

57

Transcript of Hello from CS - JAIST...

Page 1: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted
Page 2: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Hello from CS

Page 3: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Security on Cloud Computing, Query Computation and Data Mining on Encrypted Database

Professor David CHEUNG

Head, Department of Computer Science The University of Hong Kong

RIVF 2010 Hanoi

Page 4: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

What is Cloud Computing ???

Page 5: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

From Youtube

Song from Youtube

Page 6: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

What the Cloud can offer ?‐ unlimited supply of computing power 

(electricity)

Page 7: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

What do they say about Cloud Computing?

Page 8: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

• 27.5 M hits in Google Search on “cloud computing”

• “The term cloud is used as a metaphor for Internet, based on how Internet is depicted in the network diagrams” –Wikipedia

• “The ability to connect to software and data on the Internet (the cloud) instead of your hard drive or local area network” ‐ComputerWorld

Page 9: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Some examples of Cloud 

Computing?

Page 10: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Gmail – Web mail

Page 11: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Flickr – Photo Album

Page 12: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Google Docs – Internet Docs

Page 13: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Amazon Web Services –Infrastructure as a services

Page 14: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

The three pillars of Cloud 

Computing?

Page 15: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Google Docs – Internet Docs

Page 16: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Virtualization

Separation of applications and infrastructure

Page 17: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Software as a Service

Page 18: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Utility Computing

Page 19: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Impacts and Riskof Cloud 

Computing?

Page 20: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Survey by World Economic Forum (Dec 2009)

Page 21: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Barriers – from WorldEconomicsForum

• Lack of Understanding• Interoperability• Fear of vendor lock‐in• Privacy

• Security• Governance• Integration and Migration

Page 22: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

How about security inCloud 

Computing?

Page 23: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Trust me, it is very secure.

How about that 0.001% chance ???

Page 24: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Can we run query applications on Cloud if the provider cannot be 

trusted??

Secure Computation on Encrypted

Database

Page 25: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Secure Computation on Encrypted Data – What ?

X Y

Z

encryption

Computation :x+y ??

find a: a > 0 ?

data

Page 26: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Security on Cloud Computing

Amazon’s EC2database

E(DB)(U.S.)

encrypted customerdata

(submit to Cloud for query services )

A’s customers(U.K.)

Company A(Hong Kong)

customerqueries

Gartner listed SEVEN security issues in Cloud Computing:

4. Data Segregation : encryption scheme must be available to protect the corporate data

Challenge: How to perform computation (queries) on encrypted data ??

Page 27: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Computation on Encrypted Data –what is the key issue ??

Bob: the master

Alice: untrusted slave

xy…

E(x), E(y)

I want to find x+y;I want Alice to do it for me

But I don’t trust Alice

E(x)E(y)

Result = E(x) + E(y)

R = E(x) + E(y)

Privacy homomorphism (additive):

E(x) + E(y) = E(x + y)

E-1(R)

= E-1(E(x) + E(y))

= E-1(E(x + y))

= x + y

[Rivest, Adleman, Dertouzos, Foundations of Secure Computation 1978]

Can I use R to recover (x+y) ??

Keyissue??

need a nice and secure encryption

Page 28: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

What is a secure encryption scheme ?

A secure scheme :

It can stop the attacker and achieve the defined security goal ?

Security goal Adversary Model

Attack ModelWhat can the adversary do ?

e.g., he can access some data (plain text) and encrypt it

What does the adversary know ?

e.g., he knows some background knowledge

Security GoalWhat is the primary goal ?

e.g., prevent the adversary (service provider) from seeing the data

What is the additional requirement ?

e.g., prevent the adversary from finding any statistical information on the protected data

Page 29: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

A model for Secure Queries on Encrypted Database (SCONEDB)

• Encrypted DBMS (EDBMS) hosting at an untrusted service provider– Store encrypted data– Process queries

• A Three Players Game– Player 1 : Database owner  ‐ encrypt data and send them to 

the db at the service provider

– Player 2 : User of the database – issue queries to the EDBMS

– Player 3 : Attacker (service provider) – try to break into the encrypted database

Page 30: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Player 2 (queries issuer)

Player 1 (data owner) DB

ET()tK

R EQ(q)ET(t)

EQ()

D()

K

q

D(R)

EDBMS

E(DB)query

processing

Service provider: hosting the encrypted db

SCONEDB Model

CryptanalysisH DBA

Player 3 (attacker)

knowledge known to the attacker on the data

attacker’s guess in his attempt to break

the encryption

full access to E(DB)

Page 31: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Player 2

Player 1

EDBMS

E(DB)

DB

ET() EQ()

D()

t

queryprocessing

K

K q

D(R)

R EQ(q)ET(t)

CryptanalysisH DBA

Player 3 (attacker)

Database hosting at the service provider

Define an encryptionscheme (ET,EQ and D) and a queryprocessingmethod on E(DB) such that query resultsreturned are correct and the attackercannot compromisethe E(DB), i.e., DBA is empty, given background knowledge H.

Problem definition:SCONEDB Model

Page 32: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Attack Model :Three levels of background knowledge

• Basic capability: attacker has full access to encrypted data

• Background knowledge (a three level model):– Level 1 : no background knowledge

– Level 2 : attacker knows some records in DB (plain text)

– Level 3 : attacker knows some records in DB and the encrypted values of these records, i.e., knows some (x, E(x)) pairs

x2

DB

x1

x1’x2’

E(DB)

x3’E

x3 Level 1attackerLevel 2attackerLevel 3attacker

? ? ?

Page 33: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Attack model – different background knowledge level

• The attacker has full access to the EDBMS (encrypted database, encrypted queries)

• Three levels of attacker model

– Level 1: has no background knowledge (i.e., H = empty set) (basic level);

• e.g., service provider knows nothing about the business of the data owner

– Level 2: knows a set of points P in DB (practical level);• e.g., the adversary is a customer of the bank

– Level 3: knows a set of points P in DB and their corresponding encrypted values in E(DB) (avoidable);

• e.g., the adversary creates a new account yesterday and he observes there is only one new encrypted account added since yesterday

If a scheme survives a higher level attack, it survives a lower level one.

Page 34: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

SCONEDB on an important query type: Secure kNN Computation

• We develop an encryption scheme for kNN queries on SECONEDB to explore its applicability

• k nearest neighbor query (kNN)

– Database DB: a set of d dimensional points

– Given a query point q, find the k nearest points to q in the database

q

x1x2

d2d1

DB

x2 is the 1-nearest neighbor of q

Page 35: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Is Distance Preserving Transformation (DPT) an answer to the kNN problem?

• E is a DPT if– d(x, y) = d’(E(x), E(y))

• kNN can be computed on E(DB) if E is a DPT

q

x1x2

d2d1

DB

q’

x1’

x2’d2’

d1’

E(DB)

E

(DPT)

Is this a real solution? Can it survive attack?

Nice property ??????

Page 36: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

DPT fails at levels 2 and 3 attack

• DPT fails at a level 3 attack

• DPT also fails at a level 2 attack (signature attack)

y (?)

x1

x2

DB

x3

y’

x1’x2’

E(DB)

x3’

d1’

d2’

d3’

d1

d2

d3

Attacker’s background knowledge (level 3)

Attacker wants to compromise the encryption of y’

E

(DPT)

DPT is not a good solution, it is not secure enough.

Page 37: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

An Asymmetric Scalar Product Preservation Encryption 

Scheme 1: Asymmetric Scalar Product Preservation Encryption [SIGMOD 2009]

x’ = ET(x) = MT(xT, -0.5 ||x||2 )T

q’ = EQ(q) = M-1(r(qT, 1)T)

x = D(x’) = Πd((MT)-1(x’))

Encryption key: M is a (d+1) dimension random invertible matrix

Page 38: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Scheme 1: An ASPE encryption scheme

Theorem: Let x1’, x2’, and q’ be the encrypted values of x1, x2, and q with scheme 1, then 

||x2 – q|| < ||x1 – q|| iff  x2’.q’ > x1’.q’ 

and the scheme is not a DPT. (Very nice !!!!)

q

x1x2

d2d1

E

(Scheme 1)

d dimensions d+1 dimensions

x1’

x2’

q’

Page 39: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Scheme 1: how safe ?

• Scheme 1 resists level 2 attack – it preserves some nice asymmetric scalar product and it has broken the curse of distance preservation

• Unfortunately, Scheme 1 fails level 3 attack – if enough (x, ET(x)) pairs are known to the attacker, the random matrix M can be solved (compromised)

Page 40: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Scheme 2: Asymmetric Random Splitting

• Scheme is based on a splitting technique done independently and randomly on each dimension [SIGMOD 2009]

• 2d+1 possible splitting configurations for (d+1)‐dimensional case – exponentially many configurations for the adversary to guess

• Scheme 2 resists level 3 attack if the attacker cannot derive the splitting configuration 

• If the number of dimensions is over 80, the scheme is as safe as 1024‐bit RSA key 

Page 41: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

What have we shown you ? 

• It is possible to process queries on encrypted data – so, query processing on cloud service provider could be done and need new techniques 

• This is a practical approach – the security level is not as strong as the conventional goal (e.g., semantic security); but nevertheless practical and has an implementable performance 

Page 42: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Can we do data mining on Cloud if the provider cannot be 

trusted??

Data Mining on Encrypted Data

Page 43: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Example: Association rules mining

• In the form of X => Y• Meaning

– If a transaction contains itemset X, the transaction will probably contain itemset Y

• A rule must have– High support : number of transactions that include XY– High confidence : the ratio of number of transactions containing XY 

to number of transactions containing X but not Y.

• Key issue : compute large item sets– X is large if the percentage of transactions containing X is larger than a threshold [Agrawal VLDB 94]

Page 44: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Data mining by a Service Provider (cloud)

Data Mining ServiceProvider (cloud computing)DB DB’Transformer

AssociationRules’

AssociationRules

DataOwner

Send to SPEncryption

Decryption

Page 45: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Solution: Item mapping (encryption cipher)

• T is a set of transactionst = {cheese, book, bread, chocolate, ..}

• bread ‐> 54   (item mapping)• chocolate ‐> 165• t = <cheese, book, bread, chocolate> becomes t’ = <8, 69, 54, 165>  

• <54, 165> is large to the miner, but what is it ? <cheese, book> or <bread, chocolate> ???

• Similar to substitution cipher used in encryption of text

Page 46: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

A More Secure Mapping ??

• A one‐to‐n item mapping (more secure)– B: a set of items

– m: I ‐> 2B

• Example– m(a) = {1, 4, 5}

– m(b) = {2}

– m(c) = {3, 5}

Page 47: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Itemset mapping using one‐to‐n item mapping (encryption)

• m: I ‐> 2B : one‐to‐n item mapping• M: 2I ‐> 2B : itemset mapping

• Example:– M(<a, c>) = <1, 3, 4, 5>– M(<b, c>) = <2, 3, 5>– M‐1(<1, 2, 4, 5>) = <a, b>– M‐1(<1, 2, 3, 4, 5>) = <a, b, c>

m:a -> {1, 4, 5}b -> {2}c -> {3, 5}

Page 48: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Correctness – collision on one‐to‐n mapping

• <a, b>=><1, 2, 3>

• <a, b, c>=><1, 2, 3>

Collisions!

Decryption failure !

• <a>=><1, 2>

...

• <a,b>=><1, 2, 3>

• <a, b, c>=><1, 2, 3, 4>

m:a -> {1, 2}b -> {2, 3}c -> {1, 3}

m’:a -> {1, 2}b -> {2, 3}c -> {2, 4}

A good one-to-n mapping: the mapping of each item must contain unique stuff

Page 49: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

One‐to‐n vs one‐to‐one

• one‐to‐n vs one‐to‐one?– Intuitively, one‐to‐n should be more secure

Unfortunate Scenario:• one‐to‐n + item mapping

= one‐to‐one + item mappingOur solution :

– Add a random component to transaction transformation

– It will make one‐to‐n always better (more secure) than one‐to‐one [VLDB 2007]

Page 50: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Criteria for a valid transformation

• Correctness ‐ The randomly added items must not affect the decryption

• Completeness ‐ A good random transformation algorithm should generate all possible transformations to make it not easy to guess

Page 51: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Algorithm to perform valid and complete transformation

t = <…>

StartMeet

quota?

a->…b->…

…->…

Mappings

No

N(t)

Pick one

x->x1, …,xn

History

Stores items we must not add

xi, …,xj

Filter

E

E = Ø at start

Some add to E

Others to history

Next

Add E to result

Page 52: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Integrity concerns

• The mining results returned from the service provider could be wrong– Bugs in their mining program– Only compute part of answer (laziness) – Adversely modify the results  (malicious)

• Two problems– Soundness

• All rules are correct (no false positive)• All rules are with correct support and confidence

– Completeness• All correct rules are included (no false negative)

Page 53: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Light weight audit of mining result

• Pre‐processing + light‐weighted verification [VLDB 2009]

T

Data owner

T T

FI

Transformations

Service provider

FI

Audit Environment

FrequentItemsets

FIVerifications

auxiliary data

^

^

U

R

Page 54: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

What have we shown you ? 

• It is possible to do data mining on encrypted data –at least it can be done on association mining; also the mining result is not visible to the service provider

• We can audit the mining result returned by a service provider; therefore, the integrity is protected against the service provider

Page 55: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Publications on Computation and Mining on Encrypted Data

• Data mining on encrypted data– [Wong, Cheung, Hung, Liu] 

Protecting Privacy in Incremental Maintenance for Distributed Association Rule Mining, PAKDD 2008.

– [Wong, Cheung, Hung, Kao and Mamoulis] 

Security in Outsourcing of Association Rule Mining, VLDB 2007.

• Integrity of data mining on encrypted data– [Wong, Cheung, Hung, Kao and Mamoulis] 

An Audit Environment for Outsourcing of Frequent Itemset Mining, Proceeding of PVLDB 2009 (VLDB).

• Secure kNN computation on encrypted data– [Wong, Cheung, Kao and Mamoulis] 

Secure k‐NN Computation on Encrypted Databases, SIGMOD 2009.

• Privacy preservation– [Wong, Mamoulis and Cheung]

Non‐homogeneous Generalization in Privacy Preserving Data Publishing, SIGMOD 2010.

Page 56: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

Acknowledgement

• Wong Wai Kit• Ben Kao• Nikos Mamoulis• Edward Hung

Page 57: Hello from CS - JAIST 北陸先端科学技術大学院大学bao/IEEE-RIVF2010/IEEE-RIVF-DavidCheung.pdf · Security on Cloud Computing, Query Computation and Data Mining on Encrypted

The End

Thank you. Question please !!!!