International Journal of Network Security, Vol.10, No.3, PP.213{219, May 2010 213 Evaluating The Performance of Symmetric Encryption Algorithms Diaa Salama Abd Elminaam1, Hatem Mohamed Abdual Kader2, and Mohiy Mohamed Hadhoud2 (Corresponding author: Diaa Salama Abd Elminaam) Higher Technological Institute 10th of Ramadan City, Egypt1 Faculty of Computers and Information Minu¯ya University, Egypt2 (Email: fds desert, hatem6803, [email protected] (Received Feb. 16, 2009; revised and accepted May 12, 2009) Abstract bits keys [2, 3, 5, 6, 8, 13, 14, 15, 17]. The most com- mon classi¯cation of encryption techniques can be shown Internet and networks applications are growing very fast, in Figure 1. so the needs to protect such applications are increased. Encryption algorithms play a main role in information security systems. On the other side, those algorithms con- Cryptography sume a signi¯cant amount of computing resources such as CPU time, memory, and battery power. This paper pro- vides evaluation of six of the most common encryption algorithms namely: AES (Rijndael), DES, 3DES, RC2, Public-key Private-key Protocols Blow¯sh, and RC6. A comparison has been conducted for those encryption algorithms at di®erent settings for each algorithm such as di®erent sizes of data blocks, dif- RSA& Others Block Stream ferent data types, battery power consumption, di®erent key size and ¯nally encryption/decryption speed. Exper- imental results are given to demonstrate the e®ectiveness RC4& Others of each algorithm. BlowFish &Others AES RC6 Keywords: 3DES, AES, blow¯sh, computer security, DES, encryption techniques, RC2, RC6 Figure 1: Overview of the ¯eld of cryptography 1 Introduction This paper examines a method for evaluating perfor- Many encryption algorithms are widely available and used mance of selected symmetric encryption of various al- in information security [8, 9, 10]. They can be categorized gorithms. Encryption algorithms consume a signi¯cant into Symmetric (private) and Asymmetric (public) keys amount of computing resources such as CPU time, mem- encryption. In Symmetric keys encryption or secret key ory, and battery power. Battery power is subjected to encryption, only one key is used to encrypt and decrypt the problem of energy consumption due to encryption al- data. In Asymmetric keys, two keys are used; private and gorithms. Battery technology is increasing at a slower public keys. Public key is used for encryption and private rate than other technologies. This causes a \battery gap" key is used for decryption (e.g. RSA and ECC). Public [1, 12]. We need a way to make decisions about energy key encryption is based on mathematical functions, com- consumption and security to reduce the consumption of putationally intensive and is not very e±cient for small battery powered devices. mobile devices [4, 6, 14]. There are many examples of This study evaluates six di®erent encryption algo- strong and weak keys of cryptography algorithms like rithms namely; AES, DES, 3DES, RC6, Blow¯sh, and RC2, DES, 3DES, RC6, Blow¯sh, and AES. RC2 uses RC2. The performance measure of encryption schemes one 64-bit key. DES uses one 64-bits key. Triple DES will be conducted in terms of energy, changing data types (3DES) uses three 64-bits keys while AES uses various - such as text or document, Audio data and video data- (128,192,256) bits keys. Blow¯sh uses various (32-448); power consumption, changing packet size and changing default 128bits while RC6 is used various (128,192,256) key size for the selected cryptographic algorithms. International Journal of Network Security, Vol.10, No.3, PP.213{219, May 2010 214 2 Related Work scheme is calculated as the total plaintext in bytes en- crypted divided by the encryption time [18]. To give more prospective about the performance of the The CPU process time is the time that a CPU is com- compared algorithms, this section discusses the results mitted only to the particular process of calculations. It obtained from other resources. reflects the load of the CPU. The more CPU time is used It was shown in [14] that energy consumption of dif- in the encryption process, the higher is the load of the ferent common symmetric key encryptions on hand held CPU. devices. It is found that after only 600 encryptions of a 5 The CPU clock cycles are a metric, reflecting the en- MB ¯le using Triple-DES the remaining battery power is ergy consumption of the CPU while operating on encryp- 45% and subsequent encryptions are not possible as the tion operations. Each cycle of CPU will consume a small battery dies rapidly. amount of energy. It was concluded in [7] that AES is faster and more e±- The following tasks that will be performed are shown cient than other encryption algorithms. When the trans- as follows: mission of data is considered there is insigni¯cant di®er- ence in performance of di®erent symmetric key schemes ² A comparison is conducted between the results of the (most of the resources are consumed for data transmis- selected di®erent encryption and decryption schemes sion rather than computation). Even under the scenario in terms of the encryption time at two di®erent en- of data transfer it would be advisable to use AES scheme coding bases namely; hexadecimal base encoding and in case the encrypted data is stored at the other end and in base 64 encoding. decrypted multiple times. ² A study is performed on the e®ect of changing packet A study in [19] is conducted for di®erent popular secret size at power consumption during throughput for key algorithms such as DES, 3DES, AES, and Blow¯sh. each selected cryptography algorithm. They were implemented, and their performance was com- pared by encrypting input ¯les of varying contents and ² A study is performed on the e®ect of changing data sizes. The algorithms were tested on two di®erent hard- types - such as text or document, audio ¯le, and ware platforms, to compare their performance. They had video ¯le - for each cryptography selected algorithm conducted it on two di®erent machines: P-II 266 MHz on power consumption. and P-4 2.4 GHz. The results showed that Blow¯sh had a very good performance compared to other algorithms. ² A study is performed on the e®ect of changing key Also it showed that AES had a better performance than size for cryptography selected algorithm on power 3DES and DES. It also shows that 3DES has almost 1/3 consumption. throughput of DES, or in other words it needs 3 times than DES to process the same amount of data [20]. In [11] a study of security measure level has been pro- 4 Experimental Results posed for a web programming language to analyze four 4.1 Di®erentiate Output Results of En- Web browsers. This study consider of measuring the per- formances of encryption process at the programming lan- cryption (Base 64, Hexadecimal) guage's script with the Web browsers. This is followed by Experimental results are given in Figures 2 and 3 for the conducting tests Experimental in order to obtain the best selected six encryption algorithms at di®erent encoding encryption algorithm versus Web browser. method. Figure 2 shows the results at base 64 encod- ing while Figure 3 gives the results of hexadecimal base encoding. We can notice that there is no signi¯cant dif- 3 Experimental Design ference at both encoding method. The same ¯les are en- crypted by two methods; we can recognize that the two For our experiment, we use a laptop IV 2.4 GHz CPU, in curves almost give the same results. which performance data is collected. In the experiments, Time consumption of encryption algorithm (base 64 the laptop encrypts a di®erent ¯le size ranges from 321 K encoding) byte to 7.139Mega Byte139MegaBytes for text data, from 33 Kbytes to 8262 Kbytes for audio data, and from 4006 Kbytes to 5073 Kbytes for video ¯les. 4.2 E®ect of Changing Packet Size for Several performance metrics are collected: 1) Encryp- Cryptographic Algorithms on Power tion time; 2) CPU process time; and 3) CPU clock cycles Consumption and battery power. 4.2.1 Encryption of Di®erent Packet Size The encryption time is considered the time that an en- cryption algorithm takes to produce a cipher text from Encryption time is used to calculate the throughput of a plaintext. Encryption time is used to calculate the an encryption scheme. The throughput of the encryp- throughput of an encryption scheme. It indicates the tion scheme is calculated by dividing the total plaintext speed of encryption. The throughput of the encryption in Megabytes encrypted on the total encryption time for International Journal of Network Security, Vol.10, No.3, PP.213{219, May 2010 215 Figure 3: Time consumption of encryption algorithm Figure 2: Time consumption of encryption algorithm (Hexadecimal encoding) (base 64 encoding) each algorithm in. As the throughput value is increased, the power consumption of this encryption technique is de- creased. Experimental results for this compassion point are shown Figure 4 at encryption stage. The results show the superiority of Blow¯sh algorithm over other algorithms in terms of the processing time. Another point can be noticed here; that RC6 requires less time than all algo- Figure 4: Throughput of each encryption algorithm rithms except Blow¯sh. A third point can be noticed here; (Megabyte/Sec) that AES has an advantage over other 3DES, DES and RC2 in terms of time consumption and throughput. A fourth point can be noticed here; that 3DES has low per- formance in terms of power consumption and throughput 4.3 The E®ect of Changing File Type when compared with DES.