Title: Security of Mobile Banking
1Security of Mobile Banking
- Presented by
- Ming Ki Chong mchong_at_cs.uct.ac.za
- Kelvin Chikomo kchikomo_at_cs.uct.ac.za
- Supervisor
- Alapan Arnab, Andrew Hutchison
-
2Overview
- Introduction
- SMS Banking
- GPRS Banking
- Conclusion
3Introduction
4Hypothesis
- There are currently many flaws in the present
mobile banking implementations. - We believe we can build a more secure banking
implementation using both SMS and GPRS protocols
5Project Outcomes
- Developed application should abide to the
following security principles - Confidentiality
- Authenticity
- Integrity
- Non-repudiation
- Availability
- Comparison of SMS and GPRS implementations
6Timeline
7Work Division
- Ming Ki Chong
- SMS Banking
- Kelvin Chikomo
- GPRS Banking
8Work Division
GSM GPRS Architecture
GSM SMS Architecture
Secure GPRS Banking
Secure SMS Banking
Secure SMS Banking Server
Secure GPRS Banking Server
Secure Mobile Banking
9SMS Banking
10SMS Banking Overview
- Back Ground Research
- GSM Architecture
- SMS Scenarios
- Current SMS banking
- What I Propose to Research
- What I Propose to Implement
- Concerns
11GSM Architecture
MS Mobile Station BTS Base Transceiver
Station BSC Base Station Controller MSC Mobile
Switching Centre GMSC Gateway MSC SMSC Short
Message Service Centre OMC Operation and
Maintenance Centre ISC International Switching
Centre EIR Equipment Identity Centre AUC Authentic
ation Centre HLR Home Location Register VLR Visito
r Location Register
12SMS Security Flaws
SMS is stored in plain text
Short Message Entity SME
SMSC
HLR
MSC
VLR
MS
Access Authenticate
1. Msg Transfer
2. Verify Restrictions
3. Forward Short Msg
4. Submit
5. Delivery Report
6. Delivery Report
13Current Mobile Banking
- WIZZIT
- MTN Mobile Banking
- Standard Bank
- FNB
- ABSA
Use WIG (Wireless Internet Gateway)
14What I Propose to Research
- Different Protocols for SMS Banking
- Security of using SMSes to Perform Transactions
- SMS Encryption
- Authentication
- Possible Attacks
15What I propose to Implement
- Mobile Banking Application Using J2ME
- Secure SMS protocol
- SMS Banking Server
- Secure Connection between the Bank Server and the
Database
16Protocol Layers
17Concerns
- Cost
- J2ME vs. WIG
- Security vs. Performance
- Security vs. Functionality
- Hardware Platform (Compatibility)
- Usability (User Interface)
18GPRS Banking
19Overview
- GPRS architecture
- Data route
- Security implementations and shortfalls
- Bank implementations (WAP)
- Handshakes
- Authentication mechanisms (Pins Voice prints)
- Security shortfalls
- What I propose to do
20Data route
21GPRS security shortfalls
- Authentication Center (RAND, Kc, Ki, SRES)
- Denial of service attack, using the RAND value.
- Problems with the A3/A8 authentication algorithm
- Problems with A5 algorithm
- Look at note
22Bank implementations (WAP)
- Handshakes
- Authentication mechanisms (Pins Voice prints)
- Security shortfalls
23Handshakes
24Authentication mechanisms
- Secret passwords
- Voice prints
- SIM verification codes
25Security Shortfalls
- There is no end-to-end encryption between client
and bank server. - Public key cryptosystems key sizes offered by the
WTLS standard are not strong enough. - Anonymous key exchange suites offered by the
WTLS handshake are not considered secure.
26 Present implementations
My proposal implementation
27What I propose to do
- Build a WAP Gateway, that links the mobile
station to the bank Server from the GPRS network. - Either implement a Wap Browser plugin or J2ME App
that will ensure Full Mutual Authentication
during handshake protocol - The Plugin or J2ME app should also update and
maintain network settings
28If time permits
- Look into using different key sizes, and
encryption algorithms like blow fish.
29Possible hindrances
- Time could be limited
- GPRS Access Point
30Future research
- Lawful tapping
- Session ID management on Bank Server side. (In
case of abbreviated handshake)
31Conclusion
32Outcome
- Two secure mobile banking solutions.
- SMS solution
- GPRS solution
- Secure banking server
- Research Paper citing shortfalls in current
systems and our new implementation.