Inter national J our nal of P o wer Electr onics and Dri v e System (IJPEDS) V ol. 11, No. 1, March 2020, pp. 284 290 ISSN: 2088-8694, DOI: 10.11591/ijpeds.v11.i1.pp284-290 r 284 V erifiable secur e computation of linear fractional pr ogramming using certificate v alidation Nedal M. Mohammed 1 , Laman R. Sultan 2 , Ahmed A. Hamoud 3 , Santosh S. Lomte 4 1,4 Department of Computer Science, Dr . Babasaheb Ambedkar Marathw ada Uni v ersity , Aurang abad, India 2 Department of Po wer Mechanics, Basra T echnical Institute, Southern T echnical Uni v ersity , Al-Basrah, Iraq 3 Department of Mathematics, Dr . Babasaheb Ambedkar Marathw ada Uni v ersity , Aurang abad, India 1 Department of Computer Science, T aiz Uni v ersity , T aiz, Y emen. Article Inf o Article history: Recei v ed Mar 28, 2019 Re vised Jul 8, 2019 Accepted Jul 31, 2019 K eyw ords: Certificate v alidation LFP Computation outsourcing V erifiable computation V erifiable secure computation of LFP ABSTRA CT Outsourcing of scientific computations is attracting increasing attention since it enables the customers with limited computing resource and storage de vices to outsource the sophisticated computat ion w orkloads into po werful service pro viders. Ho we v er , it also comes up with some security and pri v ac y concerns and challenges, such as the input and output pri v ac y of the customers, and cheating beha viors of the cloud. Moti v ated by t hese issues, this paper focused on pri v ac y-preserving Linear Fractional Programming (LFP) as a typical and practically rele v ant case for v erifiable secure multiparty computation. W e will in v estig ate the secure and v erifiable schema with correctness guarantees, by using normal multiparty techniques to compute the result of a computation and then using v erifiable techniques only to v erify that this result w as correct. This is an open access article under the CC BY -SA license . Corresponding A uthor: Nedal M. Mohammed, Department of Computer Science, T aiz Uni v ersity , T aiz, Y emen. Email: dr .nedal.mohammed@gmail.com 1. INTR ODUCTION The po werful adv antage of cloud com pu t ing is called outsourcing, where the customers with l imited computing res ource and storage de vices can outsource the sophisticated computation w orkloads into po werful service pro viders. Despite the tremendous benefits, there are man y challenges and security concerns because the cloud serv er and customer are not in the same trusted domain, to a v oid these problems [1-4]. First, to combat the security concern is applying encryption techniques to customer’ s sensiti v e information before outsourcing to the cloud b ut still, there is a challenge ho w mak es the task of computation o v er encrypted data [5,6]. Second, no guarantee from the cloud on the quality of the computed data and results. F or instance, solving financial linear programs is useful for optimizing global profits confidentiality is important because the inputs are sensiti v e information from multiple companies b ut correctness is important because the outcome represents financial v alue. In theory , correctness and pri v ac y can be achie v ed by producing cryptographic proofs of correctness in a multi-party w ay [7,8]. In [9] The y achie v ed Correctness by replicating a computation and comparing the results this done ag ainst uncorrelated f ailure. W ithout assuming uncorrelated f ailure or trusted hardw are the correctness can be done e.g., [10] by instead producing cryptographic proofs of correctness. Also, pri v ac y J ournal homepage: http://ijpeds.iaescor e .com Evaluation Warning : The document was created with Spire.PDF for Python.
Int J Po w Elec & Dri Syst ISSN: 2088-8694 r 285 can be done when the computation achie v ed by multiple computation parties using multiparty computation protocols e.g. [11,12]. In this paper , we w ant to be sure that the resul ts are correct and with the multiple mutually distrust ing in putters, also we w ant to guarantee the pri v ac y of the inputs. W e present certificate v alidation as a general technique for achie ving v erifiable secure computation of linear fractional programming. W e use of El-Gamal encryption [13-15] by combining the computation stage and the v alidation stage rather than using e xpensi v e encryption schemes such as P aillier’ s cryptosystem. The rest of the paper is or g anized as follo ws: section 2. Sho ws v erifiable computation schema. In section 3. W e describes the system model of our proposed Protocol for pri v ac y-preserving outsourcing LFP . In section 4. W e pro vide e xperimental result analysis for the proposed schema. At last the w ork conclusion is presented in section 5. 2. VERIFIABLE COMPUT A TION V erifiable computation has been studied by plenty of researchers in v arious application s cenarios. The y researched widely ho w to v erify the correctness of computations performed by untrusted parties (without pri v ac y) [16-21]. Figure 1. System Structure of V erifiable Computation Scheme. V erifiable computation schemes are normally based on either computation comple xity theory or cryptographic algorithms. Data and computations can be outsourced to another party in order to obtain a processing result in return. Ho we v er , whether the result is right or wrong could cause a potential risk for a data processing result requester . F or outsourced data processing and computations, v erification of the computation results is a critical issue to ensure the trust of Computation-as-a-Service [22]. 3. PR O T OCOL FOR PRIV A CY -PRESER VING OUTSOURCING LINEAR FRA CTION AL PR O- GRAMMING W e present main protocol for pri v ac y-preserving outsourcing with correctness guarant ees. W e compute a solution and a so-called certificate using normal multiparty computation, and then produce V erifiable secur e computation of linear ... (Nedal M. Mohammed) Evaluation Warning : The document was created with Spire.PDF for Python.
286 r ISSN: 2088-8694 a proof that the solution is v alid with respect to the certificate using the El-Gamal-based proofs [23]. 3.1. Functions of certificates and v alidating T o ef ficiently v alidate a computation result, we use certificates. In comple xity theory , a certificate is a proof that a v alue lies in a certain set that can be v erified in polynomial time. Let S 1 ; S 2 be sets and Y S 1 . A polynomial time computable predicate [ ' S 1 S 2 ] is called a v alidating function for Y if Y = f w 2 S 1 j9 c 2 S 2 : ' ( w ; c ) g : If ' ( w ; c ) w 2 Y : In our case, a computation is gi v en by a computation function ' ( y ; a; r ) ; and a v alidating function ' ( y ; a; r ) : Here, on input x , function f computes function output r and certificate a ; v alidating function ' checks that r is a v alid output with respect to x and a . W e require that if ( a; r ) = f ( y ) , then ' ( y ; a; r ) , b ut we do not demand the con v erse: the outcome of the computation might not be unique, and might merely check that some correct solution w as found, not that it w as produced according to algorithm f . (F or instance, a square root finder may return the positi v e square root while ne g ati v e square root is also v alid.) In our case study , we use that the optimality of a solution to a LFP can be ef ficiently v alidated using a certificate. 3.2. The v erifiable multiparty computation pr otocol by certificate v alidation W e present V erifiable Multiparty computation protocol by certificate v alidation (V erMPC) protocol to compute ( a; r ) = f ( x ) , and pro v e this resul t X i is correct. W e use passi v ely secure multiparty computation protocols based on ( t; n ) Shamir sharing with n = 2 t + 1 . In these protocols, the input parties encrypt and announce their inputs, then mak es a proof of kno wledge of the corresponding plainte xt then broadcast for this encryption and proof. Ne xt, the parties pro vide the plainte xt and randomness of the encryption to the tw o computation parties who will later pro v e the result is correct. The tw o computation p a rties check if the pro vided sharing of the input is consistent with the encryptions that were broadcast for pre v enting corrupted input parties learns information about both their encrypted and their secret shared inputs, this done by encrypting their shares of the inputs then using the homomorphic property of the cryptosystem for checking correctness. Then, the actual computation tak es place in the third computation party . The tw o parties holding additi v e shares of the input Shamir -share them between all three computation parties, then the computation is performed between the three partie s. These tw o of the computation parties produce the encrypted result and pro v e its correctness [24]. The computation pa rties send their additi v e shares of the result and the randomness of their encryption shares results to the resulting party (the encryptions of the certificate and proof of correctness) [13,25-27]. The result party checks the proofs of kno wledge pro vided by the in putters computes the encrypted results from its shares and use V erify algorithm to v erify the correctness. Figure 2. System structure of v erifiable computation protocol by certificate v alidation Int J Po w Elec & Dri Syst, V ol. 11, No. 1, March 2020 : 284 290 Evaluation Warning : The document was created with Spire.PDF for Python.
Int J Po w Elec & Dri Syst ISSN: 2088-8694 r 287 3.3. Secur e and v erifiable linear fractional pr ogramming The LFP is a special class of mathematical optimization e xpressed in the follo wing s tandard form [28]: max Z = cy + dy + S.t. Ay b; B y 0 ; (1) where (1) the objecti v e function is a linear fractional function (ratio of tw o linear functions) y is an n 1 v ector of v ariables which are to be determined, c and d are n 1 column v ectors of coef ficients, and set of constraints are a system of linear equalities and inequalities (af fine constraints) A is m n matrix of coef ficients, b is m 1 column v ector of coef ficients and ; are constants. B is n n nonsingular matrix. F or ins tance, the LFP represents the problem to find x 1 ; x 2 satisfying max Z = 2 x 1 + 3 x 2 x 1 + x 2 + 1 S.t. x 1 + x 2 3 x 1 + 2 x 2 3 x 1 ; x 2 0 : T o find the optimal solution of a fractional linear program, typically an iterati v e algorithm called the simple x algorithm is used after con v ert LFP to LP [29]. max F ( y ) = 2 y 1 + 3 y 2 S.t. 4 y 1 + 4 y 2 3 4 y 1 + 5 y 2 3 y 1 ; y 2 0 : A = 4 4 4 5 ; b = 3 3 ; c = 2 3 : Each iteration in v olv es se v eral comparisons and a Gaussian elimination step, making it quite hea vy for multiparty computation. F or relati v ely small instances, passi v ely secure linear fractional programming is feasible [11], b ut acti v ely secure MPC much less so when including pre-processing. Theor em: W e pro v e that y it is optimal using the optimal solution p of the so-called dual LP maximise b p such that A p c; p 0 : Pr oof: The solutions ( y 1 q ; ; y n q ) and ( p 1 q ; ; p m q ) ( y 2 Z n ; p 2 Z m ; q 2 N + ) are both optimal if the follo wing conditions hold: q 1; p b = c y ; A y q b ; y 0 ; A T p q c ; p 0 : Also, the simple x algorithm for finding y turns out to also directly gi v e p . T o turn the abo v e criterion into a set of polynomial equations, we define the certificate to consist of bit decompositions of ( q b A y ) i ; y i ; ( q c A T p ) i ; and p i ; and pro v e that each bit decomposition b 0 ; b 1 ; ::: sums up to the correct v alue v (with equation v = b 0 + 2 b 1 + ) and contains only bits (with equations b i (1 b i ) = 0) : V erifiable secur e computation of linear ... (Nedal M. Mohammed) Evaluation Warning : The document was created with Spire.PDF for Python.
288 r ISSN: 2088-8694 4. EXPERIMENT AL RESUL T The e xperimental results are the a v erage of multiple trials . W e design numerical e xperiments to e v aluate the ef ficienc y of the mechanism. W e ran our mechanism on se v eral LFPs. W e measured the time to solv e the LFP and to compute the certificate (this depends on the LFP size, number of iterations needed, and the bit length for internal computations), the time for PolyPro v e to produce a proof, and for PolyV er to v erify it (this depends on the LFP size and bit length for the proof). Figure 3. Sho ws the performance numbers of our e xperiments. T able 1. Performance of the proposed scheme for infeasible case Problem No. of V erify Pro v e Compute Size Iterations Algorithm Algorithm Certificate m=5, n=5 4 11.450 85 110 m=20, n=20 9 79.50 200 347 m=48, n=70 25 300.340 986 1100 m=103, n=150 62 1000 4806 8500 Figure 3. T ime cost for each phase of v erifiable secure computation of LFP using certificate v alidation F or the LFPs in our e xperiments, we find that producing proof adds little o v erhead to compute the solution and that v eri fying the proof is much f aster than participating in the computation. As a consequence, for the recipient, outsourcing both guarantees correctness and sa v es t ime compared to participating in the computation. In general, one e xpects the dif ference between computing the solution and pro ving its correctness to be more pronounced for lar ger problems: indeed, both the computation and the correctness v erification scale in the size of the LFP , b ut computation additionally scales in the number of iterations needed to reach the optimal solution. This number of iterations typically gro ws with the LFP size. Ho we v er , we only found this for the biggest LFP , where pro ving is o v er f o ur times f aster than computing, for the other programs, this f actor w as around tw o. An e xplanation for t his is that also the bit length of solutions (which influences pro ving time) typically gro ws with the number of iterations. 5. CONCLUSION In this paper , we combined passi v ely secure three-party computation with El-Gamal-based proofs. W e ha v e sho wn ho w to use certificate v alidation to obtain correctness guarantees for pri v ac y-preserving outsourcing of LFP . The securi ty guarantees of our model lie in between pa ssi v e security (that does not guarantee correctness in case of acti v e att acks) and acti v e security ( that also guarantees pri v ac y in this case). F or LFP , v erifying results tak es much less time than participating in an acti v ely secure computation; in f act, it e v en tak es less time than participating in a passi v ely secure computation without an y correctness guarantees. Hence, for computations on inputs of mutually distrusting parties, pri v ac y-preserving outsourcing with correct- ness guarantees pro vides a compelling combination of correctness and pri v ac y guarantees. Int J Po w Elec & Dri Syst, V ol. 11, No. 1, March 2020 : 284 290 Evaluation Warning : The document was created with Spire.PDF for Python.
Int J Po w Elec & Dri Syst ISSN: 2088-8694 r 289 REFERENCES [1] C. Hu, A. Alhothaily , A. Alra w ais, X. Cheng, C. Sturti v ant and H. Liu, ”A secure and v ariable outsourcing scheme for matrix in v erse computation, IEEE INFOCOM , 17(4) (2017), 2304–2312. [2] N. M. Mohammed and S. S. Lomte, ”Recent adv ances on secure computations outsourcing in cloud computing, Asian J ournal of Mathematics and Computer Resear c h , 24(4) (2017), 192–205. [3] N. M. Mohammed and S. S. Lomte, ”Secure computations outsourcing of mathematical optimization and linear algebra tasks: Surv e y , International J ournal for Resear c h in Engineering Application & Mana g e- ment , (2019), 6–11. [4] K. Zhou and J . Ren, ”CASO: Cost-A w are secure outsourcing of general computational problems, IEEE T r ansactions on Services Computing , 2018. [5] C. Gentry , ”Computing arbitrary functions of encrypted data, Comm. A CM. , 53(3) (2010), 97–105. [6] N. M. Mohammed, L. R. Sultan, S. S. Lomte, ”Pri v ac y preserving outsourcing a lgorithm for tw o-point linear boundary v alue problems, Indonesian J ournal of Electrical Engineering and Computer Science , 16(2) (2019), 1065–1069. [7] W . Shen, Y . Bo, C. Xianghui, C. Y u and S. Xuemin, ”A distrib uted secure outsourcing scheme for solving linear algebraic equations in ad hoc clouds, IEEE T r ansactions on Cloud Computing , 4 (2017). [8] N. Mohammed and S. Lomte, ”Secure and ef ficient outsourcing of lar ge scale linear fractional program- ming, International Confer ence on Computing in Engineering and T ec hnolo gy (ICCET) , (2019), T o Ap- pear . [9] M. Castro and B. Lisk o v , ”Practical Byzantine f ault tolerance and proacti v e reco v ery , A CM T r ansactions on Computer Systems , 20(4) (2002), 398–461. [10] B. P arno, J. Ho well, C. Gentry and M. Rayk o v a, Pinocchio: Nearly practical v erifiable computation, In Pr oceedings of S & P . , 13, 2013. [11] P . Bogetoft, D. L. Christensen, I. Damg _ a rd, M. Geisler , T . P . Jak obsen, M. Kraig aard, J. D. Nielsen, J. B. Nielsen, K. Nielsen, J. P agter , M. I. Schw artzbach and T . T oft, ”Secure multiparty computation goes li v e, In Pr oceedings of F inancial Crypto. , 09 (2009). [12] I. Damg ˚ ard, M. K eller , E. Larraia, V . P astro, P . Scholl and N. Smart, ”Practical co v ertly secure MPC for dishonest majority - Or: breaking the SPDZ limits, In Pr oceedings of ESORICS. , 13 (2013). [13] T . El Gamal, ”A public k e y cryptosystem and a signature scheme based on discrete log arithms, IEEE T r ansactions on Information Theory , 31(4) (1985), 469–472. [14] A. Lopez-Alt, E. T romer and V . V aikuntanathan, ”On-the-w ay multiparty computation on the cloud via multi k e y fully homomorphic encryption, In Pr oceedings of ST OC. , 12 (2012), 1219–1234. [15] R. T amassia and N. T riandopoulos, ”Certification and authentication of data structures, In Pr oceedings of AMW . , 10 (2010). [16] P . Ananth, N. Chandran, V . Go yal, B. Kanukurthi and R . Ostro vsk y , ”Achie ving Pri v ac y in V erifiable Computation with Multipl e Serv ers - W ith-out FHE and without Pre-processing, In Pr oceedings of PKC. , 14 (2014). [17] S. Arora and S. Safra, ”Probabilistic checking of proofs: A ne w characterization of NP , J . A CM. , 45(1) (1998), 70–122. [18] C. Baum, I. Damg ard and C. Orlandi, ”Publicly auditable secure multi party computation, In Pr oceedings of SCN , 14 (2014). [19] R. Canetti, ”Security and composition of multi-party cryptographic protocols, J ournal of Cryptolo gy , 13:2000, 1998. [20] I. Damg ard, K. Damg ard, K. Nielsen, P . S. Nordholt and T . T oft, ”Condential benchmarking based on multiparty computation, IA CR Cryptolo gy ePrint Ar c hive , 2015:1006, 2015. [21] Y . Ejgenber g, M. F arbstein, M. Le vy and Y . Lindell, ”SCAPI: The Secure Computation Application Programming Interf ace, Cryptolo gy ePrint Ar c hive , Report , 2012/629, 2012. [22] D. Fiore, R. Gennaro and V . P astro, ”Ef ficiently v erifiable computation on encrypted data, In Pr oceedings of CCS. , 14 (2014). [23] J. Hromk o vic, ”Algorithmics for hard problems - introduction to combinatorial optimization, randomiza- tion, approximation, and heuristics, Spring er , 2001. [24] B. Schoenmak ers and M. V eeningen, ”Uni v ersally V erifiable Multiparty Computation from Threshold Homomorphic Cryptosystems, Cryptolo gy ePrint Ar c hive , Report, 2015/058, 2015. V erifiable secur e computation of linear ... (Nedal M. Mohammed) Evaluation Warning : The document was created with Spire.PDF for Python.
290 r ISSN: 2088-8694 [25] S. Goldw asser , Y . Kalai and G. Rothblum, ”Dele g ating computation: interacti v e proofs for muggles, In Pr oc. of ST OC. , 08 (2008), 113–122. [26] S. Kamara, P . Mohassel and M. Rayk o v a, ”Outsourcing multi-party computation, IA CR Cryptolo gy ePrint Ar c hive , 2011:272, 2011. [27] M. K eller , G. L. Mikk elsen and A. Rupp, ”Ef ficient threshold zero kno wledge with applications to user - centric protocols, In Pr oceedings of ICITS. 12 (2012). [28] A. Charnes and W . Cooper , ”Programming with linear fractional functions, Naval Re-sear c h Lo gistics Quarterly , (1962), 181–186. [29] EB. Bajal ino v , ”Linear -fractional programming theory , methods, applications and softw are, Spring er Science & Business Media , 2013. Int J Po w Elec & Dri Syst, V ol. 11, No. 1, March 2020 : 284 290 Evaluation Warning : The document was created with Spire.PDF for Python.