Tool AT-W32REV. 2026-A
Blockchain suitability check: does your project need a blockchain?
A blockchain is only worth it in a narrow case: several parties write to the same data set, do not trust one another and do not want a shared trusted party. Otherwise a database does the same with less effort, and the duty to erase under the GDPR (DSGVO) is not a problem. The check takes you through the decision tree of Wüst and Gervais (2018), then tests the practical side and names an alternative for every result.
- free
- no sign-up
- runs in your browser
01 / CHECK
Six questions about the project, then the practical check
Decision tree · 6 questions · Practical check · 7 questions
The check classifies your project according to a published decision path. It replaces neither architecture planning nor legal advice, for example on the GDPR.
The evaluation runs in your browser. Your answers are neither sent nor stored.
02 / DECISION TREE
The decision tree after Wüst and Gervais
In 2018, Karl Wüst (ETH Zürich) and Arthur Gervais (Imperial College London) posed the question that should come before every blockchain project: is a blockchain the right technology at all? Their answer is a flow of six questions Wüst 2018. Five of the seven outcomes read “no blockchain” or “permissioned chain”; only one leads to the open chain used by Bitcoin and Ethereum. The US standards institute NIST explicitly calls this critical view the one organisations should take (NIST IR 8202, section 8).
| No. | Question | Exit on | Result | Otherwise |
|---|---|---|---|---|
| 1 | Does a data set need to be stored permanently? | No | No blockchain and no shared database | continue with question 2 |
| 2 | Do several parties write to this data set? | No | No blockchain: a database is enough | continue with question 3 |
| 3 | Is there a body that all would trust to keep the data, and that is always available? | Yes | No blockchain: a database at the trusted party | continue with question 4 |
| 4 | Are all those allowed to write known in advance? | No | Public blockchain, such as Ethereum | continue with question 5 |
| 5 | Do the writers trust one another? | Yes | No blockchain: a shared database | continue with question 6 |
| 6 | Should outsiders be able to verify the state themselves? | Yes | Permissioned blockchain, publicly readable | Permissioned blockchain, readable by participants only |
Order and outcomes after Wüst and Gervais 2018, Fig. 1; questions in our own words.
What lies behind the questions
- Writers, not readers. What is counted is who creates or changes entries. Many readers do not yet make a case for a chain; a database with read permissions is enough for that.
- The trusted party must always be available. Only then can it check every write itself. A body that is only available now and then can at least admit the participants of a permissioned chain, much like a certification authority Wüst 2018.
- Trust does not mean friendship. Contracts, liability and audit rights are often enough to ensure nobody expects forged entries. Without a trust problem, copies of a database or a managed database with rights for each participant are enough Pedersen 2019.
- Public verifiability is a question of reading. It decides between a permissioned chain that anyone may read and one that only the participants read. Koens and Poll even regard it as purely a design question, because verifiability can also be achieved without a chain, for example through read access for auditors Koens 2018.
Why, then, a practical check as well? Koens and Poll compared 30 such decision schemes. They partly contradict one another, and most name no alternative to the blockchain Koens 2018. Pedersen, Risius and Beck add questions missing from the tree, such as whether the rules change often Pedersen 2019. And since the choice of technology can hardly be changed later Kannengießer 2020, data protection and operation belong before the decision. The check therefore names an alternative for every result and, in Part 2, asks about the points where projects fail in practice.
03 / RESULTS
All results with their alternative
“No blockchain” is not a failure but often the right answer. A chain is a database that many parties keep jointly, and that costs speed and money: for running business processes, Ethereum came out two orders of magnitude above a cloud service in one comparison Rimba 2020. That is only worth it where it cannot be done without keeping it jointly.
| Result | When | Why | Alternative |
|---|---|---|---|
| No blockchain and no shared database | Question 1 with No: no data set remains. | Without a permanent data set there is nothing a chain could secure. Data that is only passed on goes through an interface between the parties involved. | A timestamp, if proof is needed. If it should later be provable that a document existed in exactly this form at a given time, a timestamp on its hash is enough. The method has been described since 1991, and the document stays secret. Timestamp services under RFC 3161 and qualified timestamps under eIDAS offer this. |
| No blockchain: a database is enough | Question 2 with No: only one body writes. | If only one body writes, a blockchain adds no security, only lower throughput and longer waiting times. An ordinary database with a backup does more here. | Signed, chained entries, if others are to verify. If customers or auditors should be able to see that nothing was changed afterwards, the application signs every entry and chains the hashes. Any later change then shows up. If outsiders are to verify this themselves, keep the entries in a public transparency log. |
| No blockchain: a database at the trusted party | Question 3 with Yes: a trusted party is always available. | If there is a body that everyone trusts and that is always available, it checks and writes every transaction itself. That is faster and considerably cheaper than a chain: in one comparison Ethereum came out two orders of magnitude above a cloud service. | The same database with a log for everyone. The participants get read access to a change log with signed entries. That way every transaction can be traced without everyone running a copy. If the body processes personal data on behalf of others, a data processing agreement (Auftragsverarbeitungsvertrag) is required. |
| No blockchain: a shared database | Question 5 with Yes: the writers trust one another. | If all writers trust one another, a database with shared write access is the best solution. Copies of a database, or a managed database with rights for each participant, are enough. | Rights for each participant and a change log. Everyone writes under their own ID, and every change stays in the log. If a dispute does arise, it can be settled who changed what and when. |
| Permissioned blockchain, readable by participants only | Question 6 with No: known writers who distrust one another, reading for insiders only. | Known participants who do not trust one another and do not want a trusted party: permissioned chains such as Hyperledger Fabric are built for this. Only approved parties may write and read. | A shared transparency log with signed entries. At heart, permissioned chains are a shared database. An append-only log with signed entries and hashes held by each party exposes forgery as well and is easier to run. Consider this solution seriously before you set up a chain. |
| Permissioned blockchain, publicly readable | Question 6 with Yes: as above, but outsiders are to verify. | Known writers who do not trust one another, and outsiders who are to verify: then only approved participants write, and anyone may read and verify. | A public transparency log or anchoring. A publicly readable log on the pattern of Certificate Transparency, or a database whose hash is written to a public chain at regular intervals. One entry a day is enough for anyone to verify that nothing has changed since. |
| No chain of your own: signed entries in a transparency log | Permissioned according to the tree, but fewer than 4 independent operators. | A permissioned chain can withstand one dishonest operator only from four independent operators upwards. With fewer it does not keep its promise: it can then only detect forgery, not outvote it. A log with signed entries detects it too, with much less effort. | Additionally anchor it in a public chain. If you want to give outsiders proof, write the hash of the log to a public chain at regular intervals. Then nobody, not even a participant, can rewrite yesterday’s state unnoticed. |
| Public blockchain, such as Ethereum | Question 4 with No: who writes is not fixed in advance; few transactions, no rush. | If the writers are not known in advance, you need a chain that anyone can take part in. Do not build your own for this: use an existing network whose security is carried by many independent participants. Large data and costly computations belong off the chain, with only their hash on it. | A Layer 2 for cheaper transactions. Rollups bundle many transactions and write them to Ethereum together. That lowers the fee per transaction and confirms faster. |
| Public blockchain, preferably on a Layer 2 | Question 4 with No, plus many transactions or confirmation within seconds. | If the writers are not known in advance, you need a chain that anyone can take part in. For your volume or speed a Layer 2 fits: rollups bundle many transactions and write them to Ethereum together. Large data belongs off the chain, with only its hash on it. | Only the important transactions directly on Ethereum. Where a single entry needs the full security of the main chain, for example with large values, the more expensive main chain is worth it. Everything else stays on the Layer 2 or off the chain. |
The alternatives in one sentence
- A database with signed, chained entries. Every entry carries a signature and the hash of the previous one. Anyone who changes something later is caught, even if they have taken over the machine Schneier 1999.
- Transparency log. A public, append-only record on the pattern of Certificate Transparency (RFC 9162). Anyone can check that an entry is included and that nothing has been changed Laurie 2014.
- Anchoring. The database stays; its hash is written to a public chain at regular intervals. One entry a day is enough as proof for everything before it; the data itself stays alongside Eberhardt 2017. The hash of a whole batch is generated by the Merkle generator.
- Timestamp. To prove that a document existed at a given time, a timestamp on its hash is enough, under RFC 3161 or as a qualified timestamp under eIDAS. The method is older than any blockchain Haber 1991.
- Layer 2. Rollups bundle many transactions and write them to Ethereum together, which saves fees and waiting time Thibault 2022.
04 / DATA PROTECTION
GDPR duty to erase and immutability
A blockchain is built so that nothing can be changed unnoticed. The GDPR, on the other hand, requires that data can be erased or corrected. A review of 114 papers names exactly this as the main problem of public chains, along with the question of who is responsible at all Belen-Saglam 2023. The European Data Protection Board (EDPB) has given a binding classification in Guidelines 02/2025. The final version 2.0 has been adopted since 7 July 2026; the check follows it.
- Justify first, then build. Anyone processing personal data through a chain must document why it has to be a chain, what alternatives there were and why a private or permissioned one is not enough (para. 46 and Recommendation 1).
- Public chains only if public access is necessary. Otherwise access has to be restricted (para. 49, Recommendation 15).
- Data off the chain. On the chain at most a hash made with a secret key, or a commitment that reveals nothing once the data is deleted. Such hashes also count as personal data; a plain hash without a key is not enough on a public chain (paras 52 to 54, Recommendation 2). How data and computations can be moved off the chain without giving up its advantages is described by Eberhardt and Tai Eberhardt 2017.
- Technical impossibility does not count. It is no excuse for a breach (para. 50). If there is no way to delete or anonymise the data in the end, no personal data belongs on the chain (Recommendation 11).
- Wallet addresses can also be personal data if the person can be identified from them (para. 26).
Chains whose entries can be changed afterwards have been researched. But they have considerable limits in both open and permissioned chains Politou 2021. And what is once on a public chain sits on every node: in Bitcoin researchers found over 1,600 embedded files, including links to illegal content Matzutt 2018. Then there is what is not personal data at all and yet is nobody else’s business: volumes, times and business partners can often be read from the transactions Sedlmeir 2022. If the database is held by a service provider working for all participants, a data processing agreement is required.
05 / VOLUME AND SPEED
How much a chain can handle and how fast it confirms
A central system is almost always superior in throughput and waiting time, because a chain has to reach agreement on every entry first Wüst 2018. How big the gap is depends on the type of chain. The table gives the values from the studies with their year; they are orders of magnitude, not a promise for your set-up.
| Central database | Permissioned, e.g. Fabric | Ethereum | Layer 2 (rollup) | |
|---|---|---|---|---|
| Throughput | very high; Visa handled over 50,000 per second at peak in 2018 | over 3,500 per second in favourable set-ups (2018), behind databases in comparisons | about 15 per second on paper (2020), shared with all users | considerably more than the main chain, because transactions are bundled |
| Final after | immediately | under one second (2018) | about 15 minutes | seconds provisionally; final with Ethereum, withdrawal from optimistic rollups after 7 days |
| Cost per transaction | operation and backup | running several nodes | fee, varies with load | fee, usually a fraction of the main chain |
| Run by | one body | the participants, at least 4 independent | thousands of independent nodes | one operator bundles, Ethereum secures |
Sources: Wüst and Gervais 2018 (Visa, Bitcoin), Androulaki et al. 2018 and Dinh et al. 2017 (Fabric), Sedlmeir et al. 2020 (Ethereum, throughput), ethereum.org, as of October 2026 (finality), OP Stack documentation (withdrawal), Thibault et al. 2022 (rollups).
From this the check derives its thresholds. At up to 1,000 transactions a day, throughput does not matter. Up to about 100,000 a day it recommends a Layer 2 on a public chain Thibault 2022, because all users share the few transactions per second of the main chain Sedlmeir 2020. Above that it advises collecting transactions and writing only the hash of a batch. For permissioned chains with very many transactions it asks for a load test, because in measurements they stayed well below databases Dinh 2017 Androulaki 2018. The limits at 1,000 and 100,000 are our own judgement based on these figures. What a transaction on Ethereum or a Layer 2 costs today is worked out by the gas calculator with current values.
06 / OPERATION
Practical check: why blockchain projects fail
- Too few independent operators. Methods that withstand dishonest participants only work as long as fewer than a third of the nodes are affected Castro 2002. So for one dishonest operator you need 4, and for two already seven. If all nodes sit with one party, the chain is that party’s distributed database, and a consortium can even replace blocks by legitimate means (NIST IR 8202, section 7.1). At heart, permissioned chains are a shared database Koens 2018.
- Rules that change often. Smart contracts and agreed rules are hard to change. Where price formulas or approvals change often, Pedersen, Risius and Beck advise against a chain Pedersen 2019. Who decides on changes should in any case be settled beforehand (EDPB, Recommendation 8; NIST IR 8202, section 7.2).
- Data from the real world. A chain secures what was entered, not whether it is true. Wüst and Gervais show this with a refrigerated shipment: the tamper-proof temperature sensor sits in the cooled compartment, the goods in the warm hold Wüst 2018. The less the sources can be trusted, the harder this oracle problem hits Caldarelli 2020.
- Electricity. Chains without proof of work use orders of magnitude less electricity than Bitcoin, but because of the many copies more than a central system Sedlmeir 2020. According to the CCRI, Ethereum cut its electricity use by more than 99.988 per cent when it switched to proof of stake (ethereum.org, as of July 2026).
If the check results in a public chain, do not build your own but use an existing network. Whether a provider is enough or a node of your own is needed is settled by the node planner. If you are issuing a token, the MiCA check shows which obligations follow under the EU regulation. All Web3 tools are in the Web3 overview. What we develop ourselves, from the wallet connection to the interface that brings data from the chain into your software, is described under Web3 development.
07 / RESEARCH
What the research says
The questions and warnings in the check rest on 19 verified papers. Six findings carry the most important decisions; the full list with verification notes is further down.
- Most paths do not lead to a blockchain. In the tree of Wüst and Gervais only one outcome ends at the open chain Wüst 2018.
- Decision aids forget the alternatives. Of 30 schemes examined, most name none; permissioned chains are at heart shared databases Koens 2018.
- Trust through a public chain costs. Two orders of magnitude more than in the cloud, and the limit per block caps throughput Rimba 2020.
- Permissioned chains do not replace a database. In measurements they stayed far behind database systems on ordinary tasks Dinh 2017.
- The GDPR and public chains do not go well together. The right to erasure is the main problem in 114 papers Belen-Saglam 2023, and changeable chains solve it only with limitations Politou 2021.
- What does not originate on the chain, the chain cannot verify. The oracle problem weighs more heavily the less the sources are trusted Caldarelli 2020.
08 / FOR AI ASSISTANTS
Using this tool from an AI assistant
Assistants such as Claude or ChatGPT that speak the Model Context Protocol (MCP) can call this page’s calculation directly instead of rebuilding it. The server is public, needs no sign-in and stores nothing. The answer always includes the link to this page, the date of the data and the limits of the result.
- Endpoint
https://aureum-tech.com/mcp- Tool
werkzeug_blockchain_check- Transport
- Streamable HTTP, read-only
| Name | Required | Meaning |
|---|---|---|
zustand | no | Tree, question 1: Does a data set need to be stored permanently? This means a state that matters later: stock, ownership, approvals, payments, a log. A calculation, or a message that is only passed on, does not count. Values: ja = Yes; nein = No; unsicher = I don’t know. |
schreiber | no | Tree, question 2: Do several parties write to this data set? Parties are companies, authorities, clubs, private individuals or devices with different owners. Ten employees of the same company are one party. Values: ja = Yes, several parties; nein = No, only one body; unsicher = I don’t know. |
ttp | no | Tree, question 3: Is there a body that all would trust to keep the data, and that is always available? For example an association, a chamber, an authority, a trustee or a service provider acting for all. It would have to check and carry out every write itself. A body that is only available now and then can at least admit participants. Values: ja = Yes, there is one; nein = No; unsicher = I don’t know. |
bekannt | no | Tree, question 4: Are all those allowed to write known in advance? Known means: there is a list of participants, and new ones are added only after admission. If anyone may write without registering, they are unknown. Values: ja = Yes, all known; nein = No; unsicher = I don’t know. |
vertrauen | no | Tree, question 5: Do the writers trust one another? Trust here means: nobody expects another to forge, backdate or suppress entries. Contracts and liability are often enough for that. Values: ja = Yes; nein = No; unsicher = I don’t know. |
oeffentlich | no | Tree, question 6: Should outsiders be able to verify the state themselves? For example customers, supervisory authorities or the public, without being participants themselves. Values: ja = Yes; nein = No; unsicher = I don’t know. |
personen | no | Practical check: Will personal data go into the entries? That includes customer numbers, wallet addresses of private individuals and hashes of such data, if the person can be identified from them. Values: nein = No; ja = Yes; unsicher = I don’t know. |
loeschen | no | Practical check: Will entries have to be erasable or correctable later? For example duties to erase under the GDPR, retention periods that come to an end, a court order, or an error that has to disappear. Values: nein = No; ja = Yes; unsicher = I don’t know. |
menge | no | Practical check: How many writes do you expect? Every new or changed entry counts, including automatic ones from devices. An estimate is enough. Values: wenige = Up to about 1,000 a day; viele = Up to about 100,000 a day; sehr_viele = More than 100,000 a day. |
tempo | no | Practical check: How quickly must an entry be final? For example at a checkout, or for an approval that someone is waiting for. Values: sekunden = Within seconds; minuten = A quarter of an hour is fine. |
betreiber | no | Practical check: How many independent organisations would run and maintain a node of their own? A node is a server that holds a copy of the chain and checks every entry. Two servers of the same company count as one operator. Values: keine = None, or only ourselves; wenige = Two or three; viele = Four or more; unsicher = I don’t know. |
regeln | no | Practical check: Do the rules of the process change often? For example price formulas, approval levels, deadlines or who may do what. Values: selten = Rarely; oft = Often. |
herkunft | no | Practical check: Where does the data come from? What matters is whether the chain can check what is entered. Values: digital = It arises in the system itself; real = From the real world. |
Questions where an assistant should call it
- “Do we need a blockchain to trace our supply chain with five partner companies?”
- “Is a blockchain worth it for a register that only we maintain, but that customers should be able to verify?”
- “May customer data go on a blockchain if we have to delete it under the GDPR?”
- “Permissioned blockchain or Ethereum Layer 2 for billing between three municipal utilities?”
How the connection works and what MCP can do in your own business is explained on the page about MCP integrations.
09 / COMMON QUESTIONS
Common questions about blockchain and databases
When is a blockchain worth it for a business?
Only in a narrow case: several parties write to the same data set, they do not trust one another, and there is no body that all of them would trust to keep the data. If the participants are known in advance, a permissioned blockchain fits; otherwise a public one such as Ethereum. In all other cases a database is the better choice. That is the decision tree of Wüst and Gervais (2018), which the check follows.
What is the difference between a blockchain and a database?
A database is kept by one body, and everyone else has to believe what it says. A blockchain is kept jointly by many participants: each holds a copy, and a new entry only counts once the participants agree on it under fixed rules. You pay for that in speed and cost. In 2018 Bitcoin managed about seven transactions per second; a payment network such as Visa handled over 50,000 at peak.
What is a permissioned blockchain, and when do you need one?
A chain on which only approved participants can write, for example with Hyperledger Fabric. It fits when the writers are known but do not trust one another and do not want a trusted party. To withstand one dishonest operator it needs at least 4 independent operators. With fewer, a transparency log with signed entries is usually the better solution.
May personal data be stored on a blockchain?
Preferably not. In its Guidelines 02/2025 (version 2.0 of 7 July 2026), the European Data Protection Board advises storing personal data off the chain and putting on the chain at most a hash made with a secret key. Technical impossibility is no excuse for breaching the duty to erase. If there is no way to delete or anonymise the data in the end, it does not belong on the chain.
How many transactions can a blockchain handle?
Public chains only a few per second, and all users share that: in 2020 Ethereum worked out at about 15. A block there counts as final after about 15 minutes. A Layer 2 bundles transactions and manages considerably more. Permissioned chains such as Fabric reached over 3,500 per second in favourable set-ups, but fell behind databases in comparative tests.
Does a blockchain use a lot of electricity?
That depends on the method. Chains with proof of work, such as Bitcoin, use a great deal. Chains without proof of work use orders of magnitude less; according to the CCRI, Ethereum cut its electricity use by more than 99.988 per cent when it switched to proof of stake. A central system with a few copies is still more economical, because on a chain every node computes everything.
What is a transparency log, and why is it named as an alternative?
A record to which entries can only be appended. The entries are signed and linked by hashes, so any later change shows up. Certificate Transparency, which browsers use to check the certificates of websites, works on this pattern. It needs no agreement among strangers and is therefore much simpler to run than a chain, yet it exposes forgery just as well.
Where does the blockchain check stop?
It classifies your project according to a published decision path and checks the most common pitfalls. It does not replace architecture planning with load testing or legal advice on the GDPR. We do not assess whether a token or business model is economically viable. We are happy to discuss what technically fits your project in your particular case.
10 / STUDIES AND SOURCES
What the tool is based on
The reference values and notes above rest on these 19 academic papers. Each has been checked via its DOI or the publisher’s page (last on 14 November 2026): it exists, it has not been retracted, and its finding still holds for the question today.
Wüst, K.; Gervais, A. (2018): Do you Need a Blockchain?. 2018 Crypto Valley Conference on Blockchain Technology (CVCBT), pp. 45–54
Finding A flow chart with six questions: store state, several writers, an always-available trusted party, writers known, writers trustworthy, public verifiability. Only one path leads to the open blockchain, two to the permissioned one, the rest to “no blockchain”. A central database has considerably more throughput and less delay; data from the real world stays only as reliable as whoever enters it.
In the tool Part 1 of the check follows the diagram question by question and outcome by outcome; the graphic traces your path through it.
doi.org/10.1109/CVCBT.2018.00011 · Conference paper
Koens, T.; Poll, E. (2018): What Blockchain Alternative Do You Need?. Data Privacy Management, Cryptocurrencies and Blockchain Technology (DPM/CBT 2018), Lecture Notes in Computer Science, pp. 113–129
Finding An analysis of 30 decision schemes: they contradict one another, and most name no alternatives to the blockchain. Permissioned chains are at heart a shared database; where one exists, proven alternatives are usually the better choice.
In the tool Every result of the check therefore names a concrete alternative, including the results that involve a blockchain.
doi.org/10.1007/978-3-030-00305-0_9 · Conference paper
Pedersen, A. B.; Risius, M.; Beck, R. (2019): A Ten-Step Decision Path to Determine When to Use Blockchain Technologies. MIS Quarterly Executive 18 (2), pp. 99–115
Finding A decision path in ten steps, tested on a prototype for Danish shipping. Without a trust problem, copies of a central database or a managed database with rights for each participant are enough. If the rules change often, the authors advise against a blockchain, because smart contracts and coordination make changes hard.
In the tool The question about changing rules comes from step 6; the alternative “shared database with rights for each participant” from step 3.
doi.org/10.17705/2msqe.00010 · Journal article
Kannengießer, N.; Lins, S.; Dehling, T. et al. (2020): Trade-offs between Distributed Ledger Technology Characteristics. ACM Computing Surveys 53 (2), pp. 1–37
Finding An overview of the properties of distributed ledgers and the trade-offs between them. The choice of underlying technology is decisive, because it can hardly be changed afterwards.
In the tool Explains why the check asks about data protection, volume and operation before the decision and not only afterwards.
doi.org/10.1145/3379463 · Review
Dinh, T. T. A.; Wang, J.; Chen, G. et al. (2017): BLOCKBENCH: A Framework for Analyzing Private Blockchains. Proceedings of the 2017 ACM International Conference on Management of Data (SIGMOD), pp. 1085–1100
Finding A benchmarking framework for permissioned chains, applied to Ethereum, Parity and Hyperledger Fabric. On ordinary data-processing tasks, all three were still far from replacing database systems.
In the tool With very many transactions, the check asks for a load test before the decision for permissioned chains.
doi.org/10.1145/3035918.3064033 · Conference paper
Androulaki, E.; Barger, A.; Bortnikov, V. et al. (2018): Hyperledger Fabric: A Distributed Operating System for Permissioned Blockchains. Proceedings of the Thirteenth EuroSys Conference (EuroSys ’18), pp. 1–15
Finding Describes Fabric, a platform for permissioned chains without a cryptocurrency of its own, with pluggable consensus and membership via ordinary identity management. In favourable set-ups, it achieved over 3,500 transactions per second with less than one second of latency.
In the tool Example and upper limit for permissioned chains in the results and in the performance table.
doi.org/10.1145/3190508.3190538 · Conference paper
Rimba, P.; Tran, A. B.; Weber, I. et al. (2020): Quantifying the Cost of Distrust: Comparing Blockchain and Cloud Services for Business Process Execution. Information Systems Frontiers 22 (2), pp. 489–507 (online 2018)
Finding A cost comparison for running business processes on Ethereum and on a cloud service: a difference of two orders of magnitude. The limit per block also acts as an upper limit on throughput.
In the tool Carries the note that trust through a public chain costs money, and the threshold for very many transactions.
doi.org/10.1007/s10796-018-9876-1 · Journal article
Thibault, L. T.; Sarry, T.; Hafid, A. S. (2022): Blockchain Scaling Using Rollups: A Comprehensive Survey. IEEE Access 10, pp. 93039–93054
Finding An overview of rollups as a way to scale: how optimistic and zero-knowledge rollups work, implementations with strengths and limits, and a performance comparison.
In the tool Justifies the recommendation of a Layer 2 when volume or speed overwhelm the main chain.
doi.org/10.1109/ACCESS.2022.3200051 · Review
Sedlmeir, J.; Buhl, H. U.; Fridgen, G. et al. (2020): The Energy Consumption of Blockchain Technology: Beyond Myth. Business & Information Systems Engineering 62 (6), pp. 599–608
Finding Blanket statements on electricity use fall short. Chains without proof of work use orders of magnitude less, but because of the many copies more than a central system. Public chains managed only a few transactions per second on paper, Ethereum at the time about 15.
In the tool Basis of the answer on electricity use and of the note that all users share the capacity of the main chain.
doi.org/10.1007/s12599-020-00656-x · Journal article
Castro, M.; Liskov, B. (2002): Practical Byzantine Fault Tolerance and Proactive Recovery. ACM Transactions on Computer Systems 20 (4), pp. 398–461
Finding The BFT replication method tolerates nodes that behave arbitrarily wrongly, as long as fewer than a third of the copies are affected. To withstand one such node you therefore need at least four.
In the tool From this, the threshold of four independent operators for a permissioned chain.
doi.org/10.1145/571637.571640 · Journal article
Belen-Saglam, R.; Altuncu, E.; Lu, Y. et al. (2023): A systematic literature review of the tension between the GDPR and public blockchain systems. Blockchain: Research and Applications 4 (2), Art. 100129
Finding A review of 114 papers on the GDPR and public chains. Three main problems: data subject rights such as the right to be forgotten can hardly be exercised, controllers and processors are hard to identify, and which law applies is often unclear.
In the tool Carries the red warning for personal data on a publicly readable chain.
doi.org/10.1016/j.bcra.2023.100129 · Review
Politou, E.; Casino, F.; Alepis, E. et al. (2021): Blockchain Mutability: Challenges and Proposed Solutions. IEEE Transactions on Emerging Topics in Computing 9 (4), pp. 1972–1986 (online 2019)
Finding An overview of methods for changing or deleting entries in blockchains after the fact, in order to meet the right to be forgotten. The approaches have considerable limitations in permissioned and open chains alike.
In the tool Shows that “changeable chains” are no simple solution to the duty to erase.
doi.org/10.1109/TETC.2019.2949510 · Review
Matzutt, R.; Hiller, J.; Henze, M. et al. (2018): A Quantitative Analysis of the Impact of Arbitrary Blockchain Content on Bitcoin. Financial Cryptography and Data Security (FC 2018), Lecture Notes in Computer Science, pp. 420–438
Finding Every participant holds a full copy of the chain, including content that does not belong there. In Bitcoin the authors found over 1,600 embedded files, including links to illegal content, which are thereby distributed to all participants.
In the tool Explains why data that will one day have to disappear does not belong on a public chain.
doi.org/10.1007/978-3-662-58387-6_23 · Conference paper
Eberhardt, J.; Tai, S. (2017): On or Off the Blockchain? Insights on Off-Chaining Computation and Data. Service-Oriented and Cloud Computing (ESOCC 2017), Lecture Notes in Computer Science, pp. 3–15
Finding Experience from blockchain projects: computing and storing on the chain is expensive and slows things down. The authors describe how computations and data can be moved off the chain without giving up the chain’s advantages.
In the tool Basis of the design principle “data off the chain, only the hash on the chain”.
doi.org/10.1007/978-3-319-67262-5_1 · Conference paper
Caldarelli, G. (2020): Understanding the Blockchain Oracle Problem: A Call for Action. Information 11 (11), Art. 509
Finding A literature review on the oracle problem: properties such as immutability are expected of chains that do not offer them, and the path of data from outside onto the chain is underestimated. The more trustworthy the overall system, the less the problem matters.
In the tool Carries the note when the data comes from the real world.
doi.org/10.3390/info11110509 · Review
Sedlmeir, J.; Lautenschlager, J.; Fridgen, G. et al. (2022): The transparency challenge of blockchain in organizations. Electronic Markets 32 (3), pp. 1779–1794
Finding Because every copy stores and executes all transactions, chains reveal more than expected, far beyond the conflict with the right to be forgotten. There is a trade-off between protecting sensitive data and the efficiency of smart contracts.
In the tool Note on trade secrets on publicly readable chains.
doi.org/10.1007/s12525-022-00536-0 · Journal article
Laurie, B. (2014): Certificate Transparency. Communications of the ACM 57 (10), pp. 40–46
Finding Describes Certificate Transparency, a system of public, verifiable, append-only logs. It manages without a blockchain and without agreement among strangers; the logs are run by known bodies, and anyone can verify them.
In the tool The model for the alternative “transparency log” when outsiders are to verify.
doi.org/10.1145/2659897 · Journal article
Schneier, B.; Kelsey, J. (1999): Secure audit logs to support computer forensics. ACM Transactions on Information and System Security 2 (2), pp. 159–176
Finding A computationally cheap method makes log entries on an untrusted machine so secure that an attacker can neither read earlier entries nor alter or delete them unnoticed.
In the tool Basis of the alternative “database with signed, chained entries”.
doi.org/10.1145/317087.317089 · Journal article
Haber, S.; Stornetta, W. S. (1991): How to time-stamp a digital document. Journal of Cryptology 3 (2), pp. 99–111
Finding Methods for digital timestamps with which a document can be neither predated nor backdated, even if the timestamp service cooperates. The document itself stays secret; the service does not have to keep anything.
In the tool The alternative when all that has to be provable is that a document existed at a given time.
doi.org/10.1007/BF00196791 · Journal article
Standards, authorities and manufacturer data
- Guidelines 02/2025 on processing of personal data through blockchain technologies, version 2.0 of 7 July 2026 · European Data Protection Board (EDPB)
- NIST IR 8202: Blockchain Technology Overview (2018), sections 7.1 to 7.3 and 8 · National Institute of Standards and Technology
- Single-slot finality: “It takes about 15 minutes for an Ethereum block to finalize” (as of 16.10.2026) · ethereum.org
- Energy consumption of Ethereum after the switch to proof of stake (as of 28.07.2026) · ethereum.org
- Withdrawal flow: seven-day challenge period on mainnet · Optimism (OP Stack)
- RFC 9162: Certificate Transparency Version 2.0 · IETF
- RFC 3161: Time-Stamp Protocol (TSP) · IETF
- Regulation (EU) No 910/2014 (eIDAS), Arts 41 and 42: qualified electronic timestamps · EUR-Lex
11 / MORE TOOLS
Still have a question?
All tools on blockchain, smart contracts and tokens are in the Web3 overview.
- AT-W33 · Businesses
MiCA check
Does your token fall under MiCA, and what obligations follow from that? - AT-W34 · Home & business
Gas calculator
What does a transaction cost on Ethereum, Arbitrum, Optimism, Base or Polygon? - AT-W35 · Home & business
Ethereum address checker
Is this Ethereum address valid and correctly written? EIP-55 checksum and comparison of two addresses. - AT-W36 · Businesses
Selector and topic calculator
Which function selector or event topic belongs to this Solidity signature, and the other way round? - AT-W37 · Businesses
Calldata decoder
What does this transaction do? Make calldata readable before you sign. - AT-W38 · Businesses
Merkle allowlist generator
Allowlist for a mint or airdrop: Merkle root and proofs from a list of addresses.
Got your result? Send us a copy, with two sentences about the project. We look at what fits technically and then build exactly that: a database with signed entries, an anchoring or a connection to a chain.
+49 3744 4399760[email protected]Mon–Fri 8:00–17:00 · Straße der Jugend 5, 08228 Rodewisch