06.11.2014 Views

A User Centric Security Model for Tamper-Resistant Devices

A User Centric Security Model for Tamper-Resistant Devices

A User Centric Security Model for Tamper-Resistant Devices

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

7.7 Analysis of the Proposed Protocols<br />

Table 7.4: Per<strong>for</strong>mance measurement (milliseconds) of the ABPL<br />

Measures SSL TLS Kerberos<br />

ABPL<br />

C1 C2<br />

Average Time 4200 4300 4240 2484 2726<br />

Best Time NA NA NA 2243 2634<br />

Worse Time NA NA NA 2554 2945<br />

Standard Deviation NA NA NA 64.53 76.28<br />

would have enabled us to implement the TEM at the underlying operating system level.<br />

We implemented the TEM at the application level and considered that similar or better<br />

per<strong>for</strong>mance can be attained if the TEM is implemented as part of the plat<strong>for</strong>m. Because<br />

the application-level implementation of the TEM cannot have memory access to measure<br />

the hash values of the client and server applications, we generated the hash of a xed array<br />

of size 556 bytes to represent an application state. The per<strong>for</strong>mance of the hash algorithm<br />

is based on the size of the input data and in real deployment of the protocol scenario it<br />

will depend on the size of the applications. The per<strong>for</strong>mance measurements <strong>for</strong> the ABPL<br />

are listed in table 7.4.<br />

The protocols (PBP and ABP) were executed on 16-bit Java Cards, and the implementation<br />

took 9799 bytes <strong>for</strong> the PBP and 8374 bytes <strong>for</strong> the ABP. The per<strong>for</strong>mance measurements<br />

were taken from two dierent sets of 16-bit Java Cards, and an average of recorded<br />

measurements <strong>for</strong> each sets is listed in table 7.5.<br />

Table 7.5: Per<strong>for</strong>mance measurement (milliseconds) of the PBP and ABPD<br />

Measures<br />

PBP<br />

ABPD<br />

Set One (C1) Set Two (C2) Set One (C1) Set Two (C2)<br />

Average Time 4436.23 4628.35 2998.71 3091.38<br />

Best Time 4078 4235 2906 3031<br />

Worse Time 5469 5875 3922 4344<br />

Standard Deviation 127.89 133.48 96.32 117.71<br />

Note: Set One (C1) means two Java Cards that are similar to the card C1 specication. Similarly, Set<br />

Two (C2) refers to the set of C2 Java Cards.<br />

The per<strong>for</strong>mance measurements in this section are only <strong>for</strong> reference our implementation,<br />

as the actual per<strong>for</strong>mance will vary depending upon the size of the client and server applications<br />

(i.e. hash generation), and the per<strong>for</strong>mance of public key operation, symmetric<br />

encryption, and random number generation on a given smart card.<br />

185

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!