Privacy Preserving k Secure Sum Protocol

Secure Multiparty Computation (SMC) allows parties to know the result of cooperative computation while preserving privacy of individual data. Secure sum computation is an important application of SMC. In our proposed protocols parties are allowed to compute the sum while keeping their individual data secret with increased computation complexity for hacking individual data. In this paper the data of individual party is broken into a fixed number of segments. For increasing the complexity we have used the randomization technique with segmentation

💡 Research Summary

The paper addresses the classic Secure Sum problem within the framework of Secure Multiparty Computation (SMC), where a set of parties wish to compute the total of their private inputs without revealing the individual values. Existing solutions, notably the protocol by Clifton et al. (2002), rely on a single round of computation in which each party adds a random mask to its value, passes the partial sum around a ring, and finally the initiator removes the mask to obtain the true total. While simple, this approach is vulnerable: if two adjacent parties collude, they can recover the intermediate party’s entire data by subtracting the two partial sums they receive.

To mitigate this vulnerability, the authors propose the k‑Secure Sum protocol. Each party splits its private data D_i into a fixed number k of segments D_i,0 … D_i,k‑1 such that D_i = Σ_j D_i,j. The parties are arranged in a unidirectional ring, and one party (P0) is designated as the protocol initiator. The protocol proceeds in k rounds, each round handling one segment index j. In a given round, the initiator sends its segment D_0,j to the next party, which adds its own D_1,j, forwards the partial sum, and so on until the sum of the j‑th segments has circulated back to the initiator. A round counter rc, initialized to k, is decremented after each complete pass. When rc reaches zero, the initiator announces the accumulated total, which equals Σ_i D_i.

Key properties of the k‑Secure Sum protocol:

1. No random masks are used, eliminating the risk that a compromised mask reveals the whole sum.
2. Segmentation increases security: colluding neighbors can only learn a single segment per round. To reconstruct the full data of a victim, they must succeed in all k rounds, effectively raising the attack complexity by a factor of k.
3. Probabilistic security analysis: The probability that a party becomes a victim when any two neighbors collude is P₁ = 2/(n‑1) for the original protocol (n = number of parties). For the k‑Secure Sum, the probability becomes P₁_k = (2/(n‑1))^k, which decays exponentially with k. The authors illustrate this with plots showing dramatic reductions for k≥2.
4. Communication overhead: The protocol requires k full passes around the ring, increasing both message count and latency proportionally to k. This is the trade‑off for higher security.

The authors also introduce an Extended k‑Secure Sum protocol. In this variant, each round incorporates a fresh random number r_j generated by the initiator. The partial sum transmitted in round j becomes S_ij = r_j + D_i,j + previous sum. After the round completes, the initiator subtracts r_j before announcing the total. This addition of randomness further obscures individual segments, forcing a colluding pair to perform 2k operations (k to recover the segment, k to recover the random mask) to obtain a victim’s full data.

The paper outlines three operational scenarios:

• All parties honest: The protocol completes correctly; the final sum is accurate, and privacy is preserved. The overhead is k times that of Clifton’s protocol.
• Malicious initiator: The initiator may tamper with the partial sums, leading to an incorrect final total, but individual data remains concealed.
• Two adjacent malicious parties: They can infer only one segment per round. Recovering all k segments requires k independent attacks, making the attack substantially more costly.

Assumptions underlying the design include: at least three parties, each possessing computational capability, secure communication channels, no leakage of partial sums, agreement on the segment count k, and honest‑but‑curious behavior (i.e., parties follow the protocol but may try to infer extra information).

The conclusion emphasizes that the k‑Secure Sum and its extended version provide a quantifiable improvement in privacy: the probability of data leakage drops exponentially with k, and the computational effort required for an adversary grows linearly (or doubly for the extended version). The authors suggest future work such as distributing segments among different parties, dynamic adjustment of k, and formal security proofs against malicious adversaries.

Overall, the paper contributes a clear, mathematically grounded enhancement to Secure Sum protocols, offering a practical method to trade increased communication cost for significantly stronger privacy guarantees in collaborative data aggregation scenarios.