=Paper=
{{Paper
|id=Vol-2542/MOD-DLT1
|storemode=property
|title=Towards a Process-oriented Analysis of Blockchain Data (invited paper)
|pdfUrl=https://ceur-ws.org/Vol-2542/MOD-DLT1.pdf
|volume=Vol-2542
|authors=Claudio Di Ciccio
|dblpUrl=https://dblp.org/rec/conf/modellierung/Ciccio20
}}
==Towards a Process-oriented Analysis of Blockchain Data (invited paper)==
https://ceur-ws.org/Vol-2542/MOD-DLT1.pdf
Joint Proceedings of Modellierung 2020 Short, Workshop and Tools & Demo Papers
42 Int. Workshop on Conceptual Modeling for Distributed Ledger Technologies
Towards a Process-oriented Analysis of Blockchain Data
Claudio Di Ciccio1
Abstract: Blockchains sequentially store the history of transactional information, in a virtually
immutable and distributed way. Moreover, second-generation blockchains such as Ethereum are
programmable environments, and every operation invocation towards the smart contracts corresponds
to a transaction sequentially collated in the ledgers. They thus allow for the controlled enactment of
multi-party processes as well as the immutable recording of their distributed execution. Despite the
verification, tracking, and monitoring of such blockchain-enabled processes appears paramount, a
formal and implemented framework encompassing those aspects is still a mostly unexplored research
avenue. The talk revolves around the current state of the art, as well as the opportunities and challenges
that arise when it comes to conducting a process-oriented analysis on data stemming from blockchains,
from a representation and modelling perspective.
Keywords: Blockchain; Distributed ledger; Process mining; Logging logic
Blockchain-based collaborations lay the backbone of processes involving multiple partici-
pants that interact between them [Me18, Hä18]. Recently, techniques have been devised
that allow for the direct translation of business process models into Smart Contracts [Di19].
Blockchains trace the sequence of tasks carried out in the course of process executions by
the totally ordered recording (upon consensus) of transactions between involved parties.
The payload of transactions can provide further information on the tasks carried out.
Second-generation blockchain technologies such as Ethereum allow Smart Contracts to
emit events that can be captured by Distributed Applications (ÐApps). Event logs and
data parameters of the transactions can reveal notifications and execution context. They
can, thus, enable process analytics on the blockchain [vdA16, Me18]. The persistence and
immutability of those data cater for auditing endeavours on the enacted processes [JH19].
Nevertheless, understanding the behaviour and performance of blockchain-enabled processes
still requires noticeable manual labour. The way in which logs and exchanged data are
engineered is tightly bound to how the the Smart Contracts are encoded. As shown
by [Di18] the interpretation of the information stored in the blockchain is far from
trivial. We can, for instance, observe that at block 1196772 of the Ethereum public
blockchain, transaction 0x656252f3. . . reports on a call of function 0xefe73dcb on contract
0x0e6e0313. . . from account 0x1387e749. . . . By reverse-engineering the Application Binary
Interface (ABI) of the invoked Smart Contract, one can extract the function signature
(specifically, Customer_Has_a_Problem()) and speculate that the function name is the
activity label [Mü19]. However, information pertaining to process semantics such as the
1 Sapienza University of Rome, Department of Computer Science, Rome, Italy. diciccio@di.uniroma1.it
Copyright © 2020 for this paper by its authors.
Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
Process-oriented Analysis of Blockchain Data 43
running process instance to which that transaction belongs, the conditions that led to that
invocation, or the role of the sender, remain obscure. This hampers the ex-post interpretation
of the sources of information, let alone their automated analysis. The promised verification
and traceability of executed processes ends up being ad-hoc and demanding manual effort,
not so differently from what used to happen when striving to understand the behaviour of
legacy systems through their logs [OGX12].
This issue calls for the introduction of a specification language that decouples the business
logic (encoded, e.g., in Smart Contracts) from the logging logic. Preliminary ideas, exposed
in [Kl19, Mü19], show interesting results to generate XES2 event logs for process mining
from transactions stored on the blockchain through metadata descriptors. We argue, though,
that a semantically rich language for logging logic is needed, so that actions carried out
via blockchain operations are connected to the stored data in a semantically expressive
way. A promising basis to this end is given by the recent Object-Centric Behavioural
Constraints (OCBC) specification language for processes [Ar19]. However, the logging
language should not dictate how the process behaves, but define the conditions under which
logging information is stored, and how.
New opportunities and unaddressed challenges open up in this context, including the
following: from a formal perspective, the problems of satisfiability of logging specifications
and of their consistency with the original process; from a design perspective, the adoption of
aspect-oriented programming approaches to decouple business logic code from the logging
logic descriptors, and the mechanisms to grant access to (parts of) the stored information;
from an implementation perspective, the trade-off between richness, abstraction and
retrievability of data on one side, and the execution and storage costs on the other side.
Acknowledgements. The author thanks Luciano García-Bañuelos, Jan Mendling, Marco
Montali, Wil van der Aalst, and Ingo Weber for the fruitful discussions and helpful insight.
The author is also grateful to Stefan Bachhofner, Dominik Felix, Dominik Haas, and Roman
Mühlberger for their investigations and active collaboration.
References
[Ar19] Artale, A.; Calvanese, D.; Montali, M.; van der Aalst, W.M.P.: Enriching Data Models
with Behavioral Constraints. In: Ontology Makes Sense. volume 316, pp. 257–277, 2019.
[Di18] Di Ciccio, C.; Cecconi, A.; Mendling, J. et al.: Blockchain-Based Traceability of Inter-
organisational Business Processes. In: BMSD. volume 319, pp. 56–68, 2018.
[Di19] Di Ciccio, C.; Cecconi, A.; Dumas, M. et al.: Blockchain Support for Collaborative
Business Processes. Informatik Spektrum, 42:182–190, 2019.
[Hä18] Härer, Felix: Decentralized Business Process Modeling and Instance Tracking Secured by
a Blockchain. In: ECIS. p. 55, 2018.
2 http://xes-standard.org/. Accessed 20/01/2020
44 Claudio Di Ciccio
[JH19] Jans, M.; Hosseinpour, M.: How active learning and process mining can act as Continuous
Auditing catalyst. Int. J. Accounting Inf. Systems, 32:44–58, 2019.
[Kl19] Klinkmüller, C.; Ponomarev, A.; Tran, A.B. et al.: Mining Blockchain Processes: Extracting
Process Mining Data from Blockchain Applications. In: Business Process Management:
Blockchain and Central and Eastern Europe Forum. pp. 71–86, 2019.
[Me18] Mendling, J.; Weber, I.; van der Aalst, W.M.P. et al.: Blockchains for Business Process
Management - Challenges and Opportunities. ACM TMIS, 9(1):4:1–4:16, 2018.
[Mü19] Mühlberger, R.; Bachhofner, S.; Di Ciccio, C. et al.: Extracting Event Logs for Process
Mining from Data Stored on the Blockchain. In: BPM Workshops. pp. 690–703, 2019.
[OGX12] Oliner, A.J.; Ganapathi, A.; Xu, W.: Advances and challenges in log analysis. Commun.
ACM, 55(2):55–61, 2012.
[vdA16] van der Aalst, W.M.P.: Process Mining - Data Science in Action. Springer, 2016.