Public procurement must be fair and safe. Governing bodies set rules for their procurement process [1]. The public procurement system not only prevents direct purchase-related malpractices but also enables organizations to access a broader pool of suppliers and obtain competitive prices. In addition, malpractices are also possible in direct purchase due to collusion between the supplier and the procurer. The graph presented by the researcher [2] shows that the bidding process benefits the procurer by securing the desired goods and services at a price below the market price. Purchasing by sealed quotation involves submitting the bid while maintaining its confidentiality [3].
Several studies focused on bidding systems with some limits. For instance, the study provided insights about auction types and various bidding strategies [4]. The research explained the advantages of Vickery auctions [5]. This auction benefits both the bidder and the auctioneer. The highest bidder must be declared a winner, but he offered the second-highest bid. In a reverse auction, suppliers get a lower price [4].
The major problems associated with purchasing sealed quotations are security, privacy, and potential collusion. These challenges can undermine the integrity of the bidding process and affect the outcomes for all participants. Wei & Dang provided limitations on the offline bidding or sealed quotation method [6]. A sealed quotation method requires a trusted third party. However, this approach has limitations of space and time. Moreover, staff involvement might lead to the disclosure of the bidder’s identity.
In the sealed-bid method, submitted bids are submitted in sealed envelopes to the procurer. In this case, the submitted bids could remain under the procurer’s protection. The sealed birds must be opened in the presence of the bidders or their representatives on the day of opening. Moreover, a comparison chart must be prepared for convenience and to determine the winner. Possible collusion or coercion might lead one bidder to obtain information about other bidders and engage in unfair practices to win the bid. However, conveying useful information to other bidders is a fraud [7]. The scammer may submit multiple bids and decide which bid to open based on previously opened bids. These challenges can harm the integrity of the bidding process and affect the outcomes for all participants. It also creates a trust issue in the system, leading vendors to show less interest in participating. Similarly, coercion causes bid rigging [8]. Thus, the sealed bidding is always vulnerable to malpractices such as illegal acquisition and manipulation. Hence, the bidding process needs to be digitalized [6].
E-auctions are introduced and adopted to address the problems in bidding systems. E-tendering in public procurement represents a significant shift towards transparency, efficiency, and accessibility. By leveraging technology, e-tendering enhances the traditional tendering system, ensuring fair competition and reducing opportunities for corruption [9]. It also introduces convenience and reduces the cost through healthy competition. It provides a bigger pool of bidders.
The authors discussed types of e-procurement systems [10]. In a principle-based system, the procurer makes the system available to bidders for registration, bid submission, and outcome checking. The procurer approaches a third-party platform to avail themselves of these facilities.
The authors in [11] wrote about the possible risks and disadvantages of a centralized system [11]. Since the centralized systems are connected to the internet, they are also prone to attacks. An unreliable or compromised third party can introduce malpractices such as bid repudiation, bid rigging, and data leakage. Since bidders cannot see each other’s bids or others’ submission processes, there is a risk of post-submission bid manipulation and related disputes. Many electronic sealed-bid systems are susceptible to third-party interference, in which auction hosts may conspire with bidders to manipulate outcomes [12]. Previous protocols have been found to leak sensitive bidder information, such as bank account details, compromising bidder privacy [12]. Cheating, its possibility, or fear of it creates psychological barriers for bidders [5].
In this case, E-Tendering systems are contributing to securing information and maintaining privacy. In addition, this system has increased transparency, efficiency and enhanced accessibility. E-Tendering can be customized to be more user-friendly, fair, and competitive [13, 14]. However, E-Tendering also requires robust cybersecurity measures to protect sensitive data [15]. Moreover, E-Tendering also requires technical knowledge to execute complex operations, such as online registration and obtaining a digital signature. Hence, small and non-technical companies are reluctant to participate in e-tendering.
Paillier (1999) proposed the use of homomorphic encryption [16]. Homomorphic encryption supports operations on data without decryption but consumes more resources. In this approach, only encrypted bids could be submitted. This approach benefits voting and other areas where privacy is essential. However, the technique is prone to ‘Chosen Plaintext Attack’ and ‘Chosen Cyphertext Attack’ where attackers select arbitrary plaintext and cyphertext to gain access [16].
The study in [17] separates the bid collection and processing activity. Bidder submits the bid in encrypted form. The auctioneer is responsible only for collecting bids. The auctioneer does not know how to process the collected bids. The auction issuer, another player in the process, uses circuit-based logic to process the bids and determine the outcome.
The research in [18] provided a universally verifiable auction protocol that hides losing bids. They published the price list or price range, along with the corresponding public key, and asked bidders to encrypt their bids with the relevant public key. However, decryption starts with the private key of the highest-priced band. If no bid opens, we can use the private key of the next bid (lower than the current price range). Once the highest bids open, opening remaining bids is optional. It hides the identity of losing bidders. However, a bidder can only choose a bid value from the list; they cannot choose a value outside the basket. This mechanism requires multiple public-private keys. Also, key leakage is possible.
The study [19] suggested that using public key infrastructure (PKI) for bidding is a secure option, but it limits participation due to the requirement for PKI [19]. Abe & Suzuki proposed using a physical booth [20]. Depositing the bid generates no receipt; hence, no tracking is possible, and there is no coercion or collusion. The bidder has no proof of the submitted bid value. Only the higher bid value remains open, and the rest remain unopened to preserve the privacy of other bidders. However, physical presence does not seem relevant and feasible in this era. It puts geographical restrictions on participation.
The author proposed a distributed calculation task, assuming that at least a few parties would be honest [21]. This method accommodates three types of players, Auction Authority (A), Seller (S) and Bibbers (B) and uses homomorphic encryption. Initially, the bidders prepare their bids and encrypt them with A’s public key; S collects all encrypted bids. After shuffling, they handed over to A. Since the homomorphic encryption is used, A can find the first and second winning bids. The winner can then be awarded the second opportunity. It serves the purpose of fault tolerance and of avoiding possible collusion. The communication mechanism and scaling are complex and costly. It also requires the bidder to submit the bids.
The approach described by the authors provides verifiable bid secrecy [7]. The bidder must submit a bid in an encrypted form, and decryption can be performed in front of all players. No complex multiparty interactive protocols are used. It requires another mechanism, ‘Notary’, as a witness to the bidding process. It also requires trusted servers to prevent manipulation.
The research in [22] proposed a multi-attribute bidding system where each attribute has its own weight set during the Request for Quotation (RFQ) preparation phase [22]. A calculation is required to determine the score for each bid and the winner.
Similarly, the authors [23] proposed a new method based on the zero-knowledge proof and distributed ledger to avoid manipulation [23]. This approach provides verifiable secrecy to the submitted bid. Each party selects a seed from the pool, attaches a random key to it, and forwards it to another party. Another party repeats the same and uses the final key as a token. The bidder must deposit the token amount. The token amount will be refunded only to the honest bidders at the end of the process. The proposed mechanism uses cryptocurrencies as incentives for participants.
The study proposes an upgraded version of the binary veto system [3], in which bids are submitted as binary values. Bid opening has the highest significant bit first, then the next lowest, and so on. It protects the identity of losing bidders.
Markle’s tree-based mechanism [11] requires iterative computation and consumes more time and computational resources than hashing the entire file. The bid amount and password are both used to generate the hash.
The study proposed a method based on Pederson commitment and homomorphic encryption [24]. Data operations were performed without decryption. But submission of the bid in encrypted form is compulsory. Bid value secrecy was maintained through homomorphic encryption.
Another study proposed a different approach to generating a key [25]. The key is used to encrypt the bid. After encryption, the bid must be submitted.
The study applied a method using blockchain and Designated Verifier Ring Signatures (DVRS) [26]. Auctioneer’s public key is used to encrypt bids before submission. Bidders may participate without revealing their identities. However, a bid must be submitted.
The authors proposed a blockchain-based, contract-based solution [27] to store bitwise bid values. Bids are evaluated from the most significant bit to the least significant bit. The highest bid price is identified after comparison. More computational power is required than for non-binary bids and compulsory bid submissions.
The ring signature technology was used to help bidders hide their identity [6]. It also used PKI to maintain secrecy. Blockchain technology is used to avoid single-point failure, providing verifiable secrecy for identity but requiring bid submissions.
The latest research in this domain is more focused on blockchain usage to support transparency and scalability. Bidders can submit bids in the form of smart contracts or other forms [28–33]. However, a larger blockchain network consumes more energy. Also, the full node has to maintain a copy of the entire ledger.
The key components from the literature and their limitations are given Table 1.
Existing methodology and limitations
| Sr | Existing methodology | Limitations |
|---|---|---|
| 1 | Bid submission on the distributed ledger | Requires special and extensive setup. Need reasonable number of nodes to maintain the copies of ledger [29, 30] |
| 2 | Submitting the bid in the encrypted form by using PKI | The risk with this approach is possibility of key access by the vendor [34, 35] |
| 3 | Submitting the bid in any form | Possibility of data leakage while transmission. If a host is vulnerable and compromised then adversary can access the bids. |
| 4 | Possible chip level manipulation | Possible data leakage by the hardware; difficult to detect and prevent [36, 37] |
- a)
To propose a procurement system that does not require the deposition of bid information in any form during registration, but requires bid information only at the time of bid opening.
- b)
To provide a mechanism where the host and other bidders can validate the disclosed bid information for post-registration manipulation.
Section II describes the design of the proposed system and algorithms. Section III focuses on results from a simulated real-time example. The results are discussed in Section IV. Section V shows the limitations of the proposed study. The last section concluded the proposed approach and provided insights into the future work.
Innovation Metrics is shown in Table 2.
Innovation Metrics
| Sr | Contribution | Novelty | Section |
|---|---|---|---|
| 1 | Submission less bid registration | Bidder does not have to submit the bid before bid opening day | II |
| 2 | Tamper-evident audit enable system | After getting the bid document the procurement host and other bidders can verify if the bid document is altered or genuine | II |
| 3 | Peer participation for better governance | Bidders are enabled to not only observe the process but also participate by sending queries. | II |
| 4 | No need of secure storage | Since the bid submission is required only at bid opening day. And data is open to all; there is no need to purchase and maintain online secure storage to store the bids. | II |
This study proposed a mechanism that allows the Bidder to register for the bidding process without initially submitting the quotation document. The proposed system depends on Hashing and random number generation. Hashing consistently produces a fixed-length output independent of the input length [38]. However, hash calculations are reverse, but straightforward calculation (obtaining a preimage) is challenging. Therefore, this study used the SHA256 algorithm to generate the hash value. A hashing function takes any data (say D) as input and produces a string output, also known as the Hash (H(D)).
The proposed auction system has two types of players: Bidders (or participants) and the Procurement Host (PH). The Procurement host platform is used for interaction. When there is a requirement for goods or services, the authority concerned prepares the ‘Request for Quotation’ (RFQ) document with all requirements, terms and conditions, and publishes it on the host platform (see ‘1 RFQ Published’ in Figure 1).
Layout of the proposed system
The RFQ has been kept in the public domain (see ‘1 RFQ published’ in Figure 1). Bidders who wish to participate can register on the platform. After registration, a bidder can submit a quotation and calculate its hash value. It can be considered a quotation identifier. The Bidder submits the quotation identifier to host against the published RFQ (see ‘2 Bidder Registers Quotation ID’ in Figure 1). After receiving the Quotation Identifier, the host logs it and publishes it in the open domain for transparency. The PH sends challenges to the Bidder to increase accountability. The PH specifies random fragments and asks the Bidder to send the hash of those fragments. Since the Bidder does not share the actual length of the quotation document, a fragment has to be specified in the form of a percentage, e.g. beginning at 10% and ending at 29%. Two random numbers, say A and B, can be generated where 0<=A<B<=100 and used as relative specifications of a random fragment. This will allow the Bidder to keep the document’s length private. After receiving the challenge (A, B), the Bidder converts them into an absolute specification using the calculations (Equations (1) and (2)).
And
The replying Bidder fetches the exact fragment, calculates the hash, and submits the same as a reply. The host logs and publishes the response for greater transparency (see ‘3 Host challenges-Bidder Responds’ in Figure 1). The proposed bidder registration process is shown in Algorithm 1.
Input:
RFQ identifier ->Rj
Quotation Document Prepared by the Bidder Bi-> RjBi
Output:
Public logs for verification of the quotation document (RjBi)
Begin:
Bi Authenticates with the PH
Bi retrieves the RFQ information with identifier Rj.
Bi registers on the host procurement platform for participation in tendering processing.
Bi prepares the quotation document for Ri->RjBi
Bi computes hash of the prepared document RjBi->H(RjBi)
Bi registers for procurement by sharing its identity Bi, Rj, H(RjBi) to the Procurement Host (PH)
PH generates challenges (Ac, Bc) for H(RjBi), where 0<=Ac<Bc<=100
Bi fetches the specified fragment of RjBi->RjBi[Ac, Bc]
Bi calculates hash value for RjBi[Ac, Bc] ->H(RjBi[Ac, Bc])
Bi shares response to PH (H(RjBi),[Ac, Bc], H(RjBi[Ac, Bc])
->Response
PH publishes the Response in public domain
End:
The RFQ and relevant quotation ID received are available on the hosting platform, making all bidders aware of the responses to the specific RFQ. To increase transparency and the sense of participation, any bidder can challenge another bidder through the host by specifying a random segment and quotation ID (See ‘4 Another Bidder challenges’ in Figure 1). Hosts forward the challenge to the bidder. After receiving the response, the host will log it and forward it to the bidder who issued the challenge (see ‘5 Host forwards Challenges –Bidder Responds’ in Figure 1).
Algorithm 2 specifies the process to issue challenges. The concerned bidder could fetch the challenge and a response submitted to the PH for keeping in the public domain.
Input:
RFQ identifier ->Rj
Quotation Document Identifiers -> H(RjBi)
Total number of bidders participated in Rj->NB
Maximum challenges allowed->C
Output:
Challenge issued ahainst H(RjBi)->[Ac, Bc]
Begin:
Other Bidder Bo fetches ongoing procurement from the Procurement Host (PH) ->Rj
BO fetches all quotations id received against Rj-> H(RjBi), where 0<i<=NB
For each nb in NB
For each c in C
Generate Ac, Bc such that 0<=Ac<Bc<=100
Issue challenge [Ac, Bc] against H(RjBnb)
End:
At the time of bid opening, all bidders disclose their quotation file. The proposed bid opening process is shown in Algorithm 3. The registration and challenge-response data are already available with the host. After receiving the original quotation file, the host validates it to ensure the data received has not been altered. The host then shares the same file with the remaining bidders so they can verify it as well. Only valid quotations are considered for comparison, and the content will determine the winner. If the bidder shares the altered file, the Absolute specifications will differ, resulting in a different file fragment. A distinct file fragment will generate distinct hash values, allowing tampering to be detected.
Input:
Ongoing Procurement->Rj
No of bidders participated in Rj->NB
Quotation document submitted by bidders Bi->RjBi
No of challenges received against quotation id H(RjBi)->C
Challenges received against H(RjBi)->[Ac, Bc]
Output:
Qualified and disqualified list of bidders
Begin:
For each Bn where 0<n<=NB
PH fetches document RjBn from Bn
PH fetches challenges and responses about RjBn
for each [Ac, Bc] where 0<c<=C
Calculate hash of the fragment ->H(RjBn[Ac, Bc])
If H(RjBn[Ac, Bc]) is not matching with R(RjBn[Ac, Bc]):
Disqualify the bid document and related bidder
Break and go to next bidder
Publish qualified and disqualified list of bidders
End:
For N bidders, if all are permitted to send total CB challenges to remaining bidders, and PH is supposed to send CPH challenges to each bidder, then the total number of hash operations required is given as N*CB+N*(CPH).
- 1)
Bid manipulation by the bidder:
A bidder uses a bid document (say B) for the registration, but while disclosing the bid, it shares B‵, where B and B‘ are not the same.
- 2)
data manipulation by the host:
A procurement Host receives bid ID (say B), response (Say R), and actual bid (Say B) from the bidder. But it publishes H‵, R‵, B‵ in the responses to any enquiry. OR, despite receiving a bid ID, a response, and the actual bid, it does not acknowledge them.
- 3)
Corrupt Host:
A procurement Host not acknowledging receipt of the data or propagating a falsified version of it after an enquiry.
- 4)
Corrupt bidder:
A bidder that uses identical versions of data for registration and bids opening.
We have implemented Python scripts to simulate the experiments and provided them here [39]. One script simulates the PH, and another script simulates bidders. The simulation supports 100 bidders, and each bidder has TWO copies of quotations (a genuine and an altered one). The nature of the bidders (honest, dishonest, etc.) was selected randomly. Both scripts communicate through the network socket. For ease of communication, we have created a JavaScript Object Notation (JSON) based communication protocol shown in Table 3. By using the standard messages, bidders communicated with the procurement host. In our experiment, we observed that the bidders could pull data, such as ongoing procurements, received bids for a particular procurement, and received challenges/responses for a specific bid. Also, the bidder was able to push data in PH, such as bid registration details, challenges to other bidders regarding their bids, and responses to related challenges received about the bid. The PH successfully discovered various categories of the bidders.
JSON based messages for communication
| Sr | Message | From | To | Details |
|---|---|---|---|---|
| 1 | ONGOING_PROCUREMENTS_RESUEST | Bidder | PH | A bidder send enquiry about ongoing procurements |
| 2 | ONGOING_PROCUREMENTS | PH | Bidder | PH sends list of ongoing procurements |
| 3 | REGISTER_BID | Bidder | PH | The bidder is willing to participate in procurement process. Makes the bid file and submit its hash. |
| 3 | PROCUREMENT_INFO_ REQUEST | Bidder | PH | A bidder requests list of bids received for the particular procurement |
| 4 | PROCUREMENT_INFO | PH | Bidder | Procurement Host shares the details |
| 6 | PUSH_RESPONSES | Bidder | PH | Bidder pushes calculated responses for each challenge |
| 7 | PULL_CHALLENGES | Bidder | PH | Bidder enquires about challenges related submitted bid. |
| 8 | CHALLENGES | PH | Bidder | Procurement host sends the list of challenges available for particular bid of the requesting bidder. |
| 9 | BID_INFO_REQUEST | Bidder | PH | A bidder requests list of bids received and related challenges and responses. |
| 10 | BID_INFO | PH | Bidder | Procurement Host sends data of all the bids received under a specified procurement along with challenges and responses |
| 11 | PUSH_CHALLENGES | Bidder | PH | A bidder issues challenges against a specific bid registered with respect to the particular procurement. |
| 12 | DISCLOSE_BID | Bidder | PH | A bidder discloses bid document. Can either share the path or submit the file |
We implemented the experiment with 5 bidders to measure the payload. The findings are shown in Table 4. The data generated at the PH is dumped into a JSON file, and its simplified tabular version is provided as an Annexure (Table 5).
Payload generated with 5 bidders in bytes
| Sr | Message Type | Payload size |
|---|---|---|
| 1 | ONGOING_PROCUREMENTS_RESUEST | 41 |
| 2 | ONGOING_PROCUREMENTS | 90 |
| 3 | REGISTER_BID | 166 |
| 4 | PROCUREMENT_INFO_REQUEST | 82 |
| 5 | PROCUREMENT_INFO | 433 |
| 6 | PUSH_CHALLENGES | 995 |
| 7 | PULL_CHALLENGES | 169 |
| 8 | CHALLENGES | 212 |
| 9 | PUSH_RESPONSES | 580 |
| 10 | BID_INFO_REQUEST | 74 |
| 11 | BID_INFO | 2396 |
| 12 | DISCLOSE_BID | 213 |
| Total | 5451 |
Simulation outcome with 5 bidders
| Sr | Bidder | Activity | Specifications | Details |
|---|---|---|---|---|
| 1 | 1 | Registration | 0–100* | d6f435b77ed2545dfa7ba496e246ea51a6c3931c3f2a813f63830209c6abd755 |
| 2 | 2 | Registration | 0–100 | 41487307562274ea9081df3a6b1dfefa32d49fdfa19736460e24ba8b76e3c5b0 |
| 3 | 3 | Registration | 0–100 | 0db5126c7488330795bd31d346b8237c9b02655f7c0942991ad2b50488cb4fac |
| 4 | 4 | Registration | 0–100 | d3c0a4807b780d43fbc240e325992b6173c8e0b1f1112ac88d258b721d821975 |
| 5 | 5 | Registration | 0–100 | 15cc5d4f2cb7be054a46bf0a61ec93c0d9282e8dfbfe74c04f26d02e62f75b28 |
| 6 | 1 | Challenge & Response | 22–97 | a9308f146199b7ca7b2a1a1171b88e6ed445a1cbb1779fb2d561161119b57cdc |
| 7 | 1 | Challenge & Response | 77–96 | f5023e83191250752a9d49da8a2bd46656d751b42f2fc867dcabfcadd815af86 |
| 8 | 1 | Challenge & Response | 84–90 | 68cb6480628178c33ad4fd170c63df5cf36f547dc28dbbc5361f3f8326225877 |
| 9 | 1 | Challenge & Response | 27–93 | 26ec16d4bf603c147074c4e33e7862f212b9569edb617eb30c984fd47ff4eb61 |
0–100 represent hashing of the entire file.
An experiment was performed on the system with an Intel Core i5-12400 processor, 8 GB of RAM, and an NVMe SSD. Where a relatively large file of size 4.4 GB was considered [40]. First, we calculated the hash of the whole file (from 0% to 100%), then decreased it by 1% at each step. To demonstrate the worst-case scenario, the program always skips the first n% and calculates the hash of the remaining (100-n) %. The proposed approach took less than 4 seconds to process a 1 GB file (Figure 2).
Size of the fragment (MB) vs Time required (Sec) to calculate SHA-256
Another experiment was conducted to randomly select a fragment of the same file and calculate its SHA-256 value. The experiment was run 1000 times. and the average data volume processed per second for SHA-256 was 281 MB.
The implemented case study shows that the proposed system does not cause any inconvenience.
We have checked the feasibility of a brute-force attack using blank files (MS Word and PDF). It was observed that issuing a challenge for 1% of the fragment makes it difficult to obtain a preimage of the hash. The possibility of a brute-force attack with several attributes is demonstrated in Table 6.
Brute force possibility check
| Sr | Attribute | Blank .docx | Blank .pdf |
|---|---|---|---|
| 1 | Size in bytes | 13,536 | 24,827 |
| 2 | 1% of size in bytes | 135 | 248 |
| 3 | 1% of size in bits (‘2’ × 8) | 1080 | 1984 |
| 4 | Possible combinations | 21080 | 21984 |
| 5 | Changing the base to 10 | 1.29 × 10325 | 1.75 × 10597 |
Even a password of more than 15 characters can take a relatively long time to crack [41]. And the attacker needs to repeat the procedure for the remaining 99 blocks. The experiment was conducted on blank files, and it was observed that the real quotation document occupies more disk space, making a brute-force attack impossible.
The identity and quotation registration are handled by separate servers, as shown in Figure 3. In this regard, the bidder used a separate channel to identify and register bids. Thus, anonymity can be provided to bidders. However, the opening quotation does not disclose the bidder’s identity. To claim the quotation, the respective bidder must share the identity file registered earlier. However, others can validate it using the logs from the identity registration host. The proposed method is platform independent and can be applied via social media or an email system.
Separate identity and quotation registration
In the proposed system, PH keeps all communication logs available in an open domain. To add a layer of accountability, this system can be integrated with email. Where a bidder can communicate with the PH through email APIs such as Gmail, Microsoft Graph, Nylas, and others [42, 43]. Email communication is admissible as valid legal evidence in Indian courts under the Information Technology Act, 2000, and the Bharatiya Saksha Adhiniyam, 2023. This can be used to detect a manipulated host.
Such a bidder can be identified and blocked by the PH.
Absolute specifications calculated by the PH or others may differ, leading to different fragments. The same input fragment produces the same SHA-256 value. In this case, the bidder can be identified and blocked. In the current scenario, there is a provision to seize Earnest Money Deposit(EMD) in case of malpractice and dishonouring the contract after winning the bid. Generally, EMD is charged 5% of the total value.
Unidentical quotations produce different SHA-256 values, enabling the detection and blocking of bidders. Manipulation detection is independent of the size of the manipulation. The manipulation detection rate can always be 100%. The time required to detect depends on the size of the quotation document and the computer’s capacity.
In this case, existing error-detection and correction technologies (like SSH or SSL) can handle accidental manipulation and protect communications.
As per the experiments [39,40], processing a 100 MB quotation document takes 0.35 seconds. Also, the communication payload would be less than 1922 bytes, requiring less than 0.5 seconds for transmission. Overall, the activity will take less than 0.85 seconds.
In India, ‘Category A’ documents have to be preserved forever (perpetuity) by the government, public sector undertakings (PSUs) as a “Permanent” record. For transparency in financial statements, the meeting minutes must be preserved permanently. After receiving the bids, they are open to the technical committee and the purchase committee. Competitors can also attend the meeting to ensure transparency. Bids are awarded, and the procedure is minimised. Bid participants cannot request the erasure of their data, regardless of the bid outcome. Hence, the bidder cannot exercise the ‘right to be forgotten’.
Timely registration, responding to challenges and disclosing the document are key phases of the proposed system. The existing systems require that the document be deposited before the deadline. There is no need for iterative communication, i.e., challenge-and-response. Since the proposed system requires only registration, not document deposition, it may lead to a scenario in which participants register the claims but do not participate in the challenge-response. There might also be a scenario in which participant(s) can join the challenge-response but could not disclose their document after the given date. Both scenarios can be addressed by charging the security deposit or participation deposit at the registration phase and freezing the deposit AND/OR blocking the bidder from future bidding processes in case of non-compliance. Moreover, these two provisions are already practiced in the existing system.
In this paper, we have proposed a novel approach for e-tendering systems that comply with existing requirements and regulations while providing greater transparency. The proposed system is based on a ‘verifiable claim without disclosure’ and ‘post-disclosure verification’. The proposed system does not require the document deposition. Therefore, the data leakage problem has been eliminated in the proposed system. Participation in the challenge-response and post-disclosure verification approach ensures the transparency and integrity of the document claimed during registration. Registration without document submission eliminates the requirement of secure storage. A bidder can keep their quotation on one isolated computer and use a network computer to interact with the host. It is concluded that run calculations only on isolated computers to obtain the quotation and the responses to the challenges received, and enter the output on a networked computer. This can provide additional confidence in the bidder’s data safety for the proposed system. In the future, we will present a separate identity host alongside the procurement host, allowing participants to remain anonymous.