I am Parwat Singh Anjana, an Applied Researcher at
Supra,
where I work on the design and analysis of high-performance blockchain
execution systems. My work primarily focuses on parallel and distributed
systems, especially blockchain execution, including conflict-aware parallel
execution (for EVM, MoveVM, and SVM), transactional memory, optimistic
concurrency techniques, performance analysis, and the foundations of
scalable transaction processing for modern decentralized platforms.
My broader interests include distributed systems, parallel computing,
transactional memory, CRDTs, and systems for data consistency and
scalability. I enjoy building rigorous models, conducting empirical
analyses, and developing practical execution engines that maximize
throughput while preserving correctness.
[2025] Supra’s parallel execution – currently the
fastest EVM executor in the world. The
$1M open challenge and SupraBTM (iBTM) were covered by major crypto and
financial news outlets, including
Business Insider,
ETF.com,
ZyCrypto,
KuCoin,
Phemex,
Ainvest,
CoinStats, and others.
More details are available in the
SupraEVM Beta documentation.
We also released a comparative analysis where SupraBTM significantly
outperforms competing engines:
Supra v/s. Monad.
[2025] Paper titled
"Block Transactional Memory: A Complexity Study"
accepted at SSS 2025 as a full paper.
[2025] Paper titled
"Efficient Parallel Execution of Blockchain Transactions Leveraging Conflict Specifications"
accepted at AFT 2025 as a full paper.
[Video]
Anjana Parwat Singh,
Sweta Kumari, Sathya Peri, Sachin Rathor, and Archit Somani,
OptSmart: A Space Efficient Optimistic Concurrent Execution of Smart Contracts,
Distributed and Parallel Databases (DAPD),
Volume 42, Number 2, pages 245–297, 2024,
ISSN: 1573-7578, Springer Nature Switzerland AG.
@article{Anjana2024OptSmart,
author = {Parwat Singh Anjana and Sweta Kumari and Sathya Peri and Sachin Rathor and Archit Somani},
title = {OptSmart: A Space Efficient Optimistic Concurrent Execution of Smart Contracts},
journal = {Distributed and Parallel Databases},
volume = {42},
number = {2},
pages = {245--297},
year = {2024},
doi = {10.1007/s10619-022-07412-y},
url = {https://link.springer.com/article/10.1007/s10619-022-07412-y}
}
Popular blockchains such as Ethereum and several others execute complex transactions in the block
through user-defined scripts known as smart contracts. Serial execution of smart contract transactions/
atomic units (AUs) fails to harness the multiprocessing power offered by modern multi-core processors.
By adding concurrency to the execution of AUs, we can achieve better efficiency and higher throughput.
In this paper, we develop a concurrent miner that proposes a block by executing AUs concurrently using
optimistic Software Transactional Memory systems (STMs). It efficiently captures independent AUs in the
concurrent bin and dependent AUs in the block graph (BG). Later, we propose a concurrent validator that
re-executes the same AUs concurrently and deterministically using the concurrent bin followed by the BG
given by the miner to verify the block. We rigorously prove the correctness of concurrent execution of AUs.
Performance benchmarks show that the
average speedup for the optimized concurrent miner is 5.21×, with a maximum of 14.96× over the serial miner.
The optimized validator achieves an average speedup of 8.61× and a maximum of 14.65× over the serial validator.
The proposed miner outperforms state-of-the-art concurrent miners by 1.02× to 1.18×, and the proposed validator
outperforms concurrent validators by 1× to 4.46×. Moreover, the optimized BG saves an average of 2.29× more
block space compared with the state-of-the-art.
Shrey Baheti,
Anjana Parwat Singh,
Sathya Peri, and Yogesh Simmhan,
DiPETrans: A Framework for Distributed Parallel Execution of Transactions of Blocks in Blockchain,
Concurrency and Computation: Practice and Experience (CCPE),
Volume 34, No. 10, Pages: e6804, 2022, ISSN: 1532-0634,
Wiley Press, New York, USA.
[*All authors contributed equally.]
@article{Baheti2019DiPETrans,
author = {Shrey Baheti and Parwat Singh Anjana and Sathya Peri and Yogesh Simmhan},
title = {DiPETrans: A Framework for Distributed Parallel Execution of Transactions of Blocks in Blockchain},
journal = {Concurrency and Computation: Practice and Experience},
volume = {n/a},
year = {2022},
pages = {e6804},
doi = {https://doi.org/10.1002/cpe.6804},
url = {https://onlinelibrary.wiley.com/doi/abs/10.1002/cpe.6804}
}
Contemporary blockchains such as Bitcoin and Ethereum execute transactions serially by miners and validators
to determine the Proof-of-Work (PoW). Such serial execution is unable to exploit modern multi-core resources
efficiently, limiting system throughput and increasing transaction acceptance latency.
The objective of this work is to increase transaction throughput by introducing parallel transaction execution using
a static analysis technique. We propose a framework, DiPETrans, for the distributed execution of the transactions
in a block. Peers in the blockchain network form a community
to execute the transactions and find the PoW in parallel using a leader–follower approach.
During mining, the leader statically analyzes the transactions, creates different groups (shards)
of independent transactions, and distributes them to followers to execute concurrently. After execution,
the community's compute power is used to solve the PoW concurrently. Once a block is created,
the leader broadcasts the block to other peers for validation.
Upon receiving a block, validators re-execute the block transactions and accept the block
if they reach the same state as shared by the miner.
Validation can also be performed in parallel using the same leader–follower strategy.
We report experiments using over 5 million real Ethereum transactions and execute them via
the DiPETrans framework to empirically validate the benefits
over traditional sequential execution. We achieve a maximum speedup of 2.2× for the miner and
2.0× for the validator with 100–500 transactions per block.
Additionally, we achieve a peak of 5× end-to-end block creation speedup using
a parallel miner over a serial miner when using six machines in the community.
Conference Papers
Parwat Singh Anjana,
Matin Amini, Rohit Kapoor, Rahul Parmar, Raghavendra Ramesh,
Srivatsan Ravi, and Joshua Tobkin,
Efficient Parallel Execution of Blockchain Transactions Leveraging Conflict Specifications,
In 7th Conference on Advances in Financial Technologies
(AFT 2025), Leibniz International Proceedings in Informatics (LIPIcs),
Volume 354, Dagstuhl, Germany, 2025.
@InProceedings{anjana_et_al:LIPIcs.AFT.2025.29,
author = {Parwat Singh Anjana and Matin Amini and Rohit Kapoor and Rahul Parmar and Raghavendra Ramesh and Srivatsan Ravi and Joshua Tobkin},
title = {Efficient Parallel Execution of Blockchain Transactions Leveraging Conflict Specifications},
booktitle = {7th Conference on Advances in Financial Technologies (AFT 2025)},
series = {Leibniz International Proceedings in Informatics (LIPIcs)},
volume = {354},
pages = {29:1--29:26},
publisher = {Schloss Dagstuhl--Leibniz-Zentrum f{\"u}r Informatik},
year = {2025},
doi = {10.4230/LIPIcs.AFT.2025.29}
}
Traditional software transactional memory (STM) has been identified as a possible abstraction for the
concurrent execution of transactions within a block, with Block-STM pioneering its application for efficient
blockchain transaction processing on multicore validator nodes.
This paper presents a parallel execution methodology that leverages conflict specification information of
the transactions for block transactional memory (BTM) algorithms. Our experimental analysis, conducted over
synthetic transactional workloads and real-world blocks, demonstrates that BTMs leveraging conflict
specifications outperform their plain counterparts on both EVM and MoveVM.
Our proposed BTM implementations achieve up to 1.75× speedup over sequential execution and outperform the
state-of-the-art Parallel-EVM (PEVM) execution by up to 1.33× across synthetic workloads.
Parwat Singh Anjana,
Srivatsan Ravi,
Block Transactional Memory: A Complexity Study,
In 27th International Symposium on Stabilization, Safety, and Security of Distributed Systems
(SSS 2025), Lecture Notes in Computer Science, Springer, 2025.
@InProceedings{Anjana2025BTMSSS,
author = {Parwat Singh Anjana and Srivatsan Ravi},
title = {Block Transactional Memory: A Complexity Study},
booktitle = {Stabilization, Safety, and Security of Distributed Systems (SSS 2025)},
series = {Lecture Notes in Computer Science},
publisher = {Springer},
year = {2025},
doi = {10.1007/978-3-032-11127-2_6}
}
Traditional software transactional memory has been identified as a possible abstraction for concurrent execution
of transactions within a block. However, in the smart contract setting, the execution order must follow the preset
order of the transactions in the block.
This paper presents a complexity study of this model for in-memory transactions (or block transactional memory)
and investigates the fundamental lower bounds that might exist. We ask the question: Do we need to maintain multiple
versions of the values associated with each account?
To answer this, we formalize the model of block transactional memory and identify the safety property that is needed.
We show that, under natural restrictions of liveness, it is necessary for smart contract transactions to maintain a
large number of versions for each account or allow read-only transactions to write to shared memory.
We then present several algorithmic designs for single-version and multi-version block transactional implementations
that provide preset serializability.
Manaswini Piduguralla,
Saheli Chakraborty,
Parwat Singh Anjana,
and Sathya Peri,
DAG-Based Efficient Parallel Scheduler for Blockchains: Hyperledger Sawtooth as a Case Study,
In Euro-Par 2023: Parallel Processing (Euro-Par 2023),
Lecture Notes in Computer Science, Volume 14100, pages 184–198,
Springer, 2023.
@InProceedings{Piduguralla2023Sawtooth,
author = {Manaswini Piduguralla and Saheli Chakraborty and Parwat Singh Anjana and Sathya Peri},
title = {DAG-Based Efficient Parallel Scheduler for Blockchains: Hyperledger Sawtooth as a Case Study},
booktitle = {Euro-Par 2023: Parallel Processing},
series = {Lecture Notes in Computer Science},
volume = {14100},
pages = {184--198},
publisher = {Springer},
year = {2023},
doi = {10.1007/978-3-031-39698-4_13}
}
We propose a parallel directed acyclic graph (DAG) based scheduler module for concurrent execution of SCTs.
This module can be seamlessly integrated into the blockchain framework, enabling efficient execution of SCTs
in a block and improving throughput.
The dependencies among the SCTs are represented as a DAG, which facilitates parallel execution.
Furthermore, the DAG is shared with block validators, conserving network-wide resources for DAG construction.
To ensure secure parallel execution, we design a secure validator capable of validating and identifying incorrect
DAGs shared by malicious block producers.
For evaluation, our framework is implemented in Hyperledger Sawtooth V1.2.6. We measure performance across multiple
smart contract applications for various schedulers. Our proposed executor exhibits a 1.58× average performance
improvement over serial execution.
Parwat Singh Anjana,
Adithya Rajesh Chandrassery,
and Sathya Peri,
An Efficient Approach to Move Elements in a Distributed Geo-Replicated Tree,
In 15th IEEE International Conference on Cloud Computing
(IEEE CLOUD 2022), pages 479–488, 2022.
@InProceedings{Anjana2022CloudMoveTree,
author = {Parwat Singh Anjana and Adithya Rajesh Chandrassery and Sathya Peri},
title = {An Efficient Approach to Move Elements in a Distributed Geo-Replicated Tree},
booktitle = {2022 IEEE 15th International Conference on Cloud Computing (CLOUD)},
pages = {479--488},
publisher = {IEEE},
year = {2022},
doi = {10.1109/CLOUD55607.2022.00071}
}
Replicated tree data structures are extensively used in collaborative applications and distributed file systems,
where clients often perform move operations. Local move operations at different replicas may be safe;
however, remote move operations may not be safe.
When clients perform arbitrary move operations concurrently on different replicas, various bugs can occur, making this
operation challenging to implement. Prior work has revealed issues such as data duplication and cycles in replicated trees.
In this paper, we present an efficient algorithm to perform move operations on a distributed replicated tree while ensuring
eventual consistency. The proposed technique focuses on resolving conflicts efficiently, requires no interaction between
replicas, and functions effectively in the presence of network partitions.
We use last-write-wins semantics for conflict resolution based on globally unique timestamps of operations.
The algorithm requires only one compensation operation to prevent cycles when move operations are applied.
Our approach achieves an effective speedup of 14.6× to 68.19× over the state-of-the-art approach in geo-replicated settings.
Sinchan Sengupta,
Sathya Peri,
and Parwat Singh Anjana,
A Self-stabilizing Minimum Average Stretch Spanning Tree Construction,
In International Conference on Networked Systems (NETYS 2022),
Lecture Notes in Computer Science, Volume 13464,
pages 119–135, Springer, 2022.
@InProceedings{Sengupta2022LowStretch,
author = {Sinchan Sengupta and Sathya Peri and Parwat Singh Anjana},
title = {A Self-stabilizing Minimum Average Stretch Spanning Tree Construction},
booktitle = {Networked Systems (NETYS 2022)},
series = {Lecture Notes in Computer Science},
volume = {13464},
pages = {119--135},
publisher = {Springer},
year = {2022},
doi = {10.1007/978-3-031-17436-0_9}
}
Stretch is a metric in the construction of spanning trees that measures the deviation in the distance between
a pair of nodes in the tree compared to its shortest distance in the underlying graph. This paper proposes a
silent self-stabilizing low-stretch spanning tree construction protocol, BuildTree, based on a
Low Diameter Decomposition (LDD) technique.
The LDD decomposes the graph into a small number of connected blocks or clusters, each with low diameter.
The proposed BuildTree algorithm generates a spanning tree with an average stretch of nO(1) and
converges to a correct configuration in O(n + Δ·η) rounds, where n is the number of nodes, Δ is the maximum cluster size,
and η is the number of clusters.
To the best of our knowledge, this is the first known use of self-stabilization to make low-stretch tree
constructions fault-tolerant.
Anjana Parwat Singh,
Hagit Attiya,
Sweta Kumari,
Sathya Peri,
and Archit Somani,
Efficient Concurrent Execution of Smart Contracts in Blockchains using Object-based Transactional Memory,
In 8th International Conference on Networked Systems (NETYS 2020),
Marrakech, Morocco, June 2020.
@InProceedings{Anjana2020netysObjSC,
author = {Anjana, Parwat Singh and Attiya, Hagit and Kumari, Sweta and Peri, Sathya and Somani, Archit},
title = {Efficient Concurrent Execution of Smart Contracts in Blockchains Using Object-Based Transactional Memory},
booktitle = {In 8th International Conference on Networked Systems (NETYS)},
year = {2021},
publisher = {Springer International Publishing},
address = {Cham},
pages = {77--93},
isbn = {978-3-030-67087-0}
}
Several popular blockchains such as Ethereum execute complex transactions through user-defined scripts.
A block typically consists of multiple smart contract transactions (SCTs). Miners execute these SCTs to append a block
to the chain, while validators re-execute them during consensus to ensure correctness. Miners receive incentives for
successfully adding valid blocks.
When executing SCTs sequentially, miners and validators fail to utilize multicore CPUs effectively, reducing throughput.
By leveraging multiple threads, higher efficiency and throughput can be achieved.
Prior work explored concurrent execution using Read-Write Software Transactional Memory Systems (RWSTMs).
Object-based STMs (OSTMs), operating on higher-level objects such as lists or hash tables, achieve even better throughput.
Multi-Version OSTMs (MVOSTMs) further increase concurrency by maintaining multiple versions per shared item.
This paper proposes an efficient framework for concurrent execution of SCTs using object semantics via optimistic
SVOSTMs and MVOSTMs. A multithreaded miner constructs a Block Graph (BG), capturing object-conflict relations between
SCTs, and embeds it into the block. Validators then re-execute SCTs deterministically using the BG.
A malicious miner may manipulate the BG (e.g., to enable double spending). To defend against such behavior, we present
a Smart Multi-threaded Validator (SMV) that detects incorrect BGs.
Experimental results show that the proposed miner and validator significantly outperform state-of-the-art SCT
execution frameworks.
Anjana Parwat Singh,
Sweta Kumari,
Sathya Peri,
Sachin Rathor,
and Archit Somani,
An Efficient Framework for Optimistic Concurrent Execution of Smart Contracts,
In 27th Euromicro International Conference on Parallel, Distributed and Network-Based
Processing (PDP 2019), Pavai, Italy, February 2019.
@INPROCEEDINGS{Anjana2019pdpEff,
author = {Anjana, Parwat Singh and Kumari, Sweta and Peri, Sathya and Rathor, Sachin and Somani, Archit},
booktitle = {2019 27th Euromicro International Conference on Parallel, Distributed and Network-Based Processing (PDP)},
title = {An Efficient Framework for Optimistic Concurrent Execution of Smart Contracts},
year = {2019},
pages = {83--92},
doi = {10.1109/EMPDP.2019.8671637}
}
Blockchain platforms such as Ethereum execute complex transactions in blocks through user-defined scripts
known as smart contracts. A block typically contains multiple such transactions executed by a miner. Validators
later re-execute these transactions as part of consensus to ensure the block is correct before it is appended
to the blockchain. Miners receive incentives for successfully proposing valid blocks.
Current blockchains execute smart contract transactions sequentially, which underutilizes multicore processors
and results in poor throughput. Adding concurrency can significantly improve performance.
This paper presents an efficient framework for concurrent execution of smart contract transactions using
optimistic Software Transactional Memory (STM) systems. Miners execute transactions concurrently using
multithreading, with STM ensuring atomicity and handling synchronization conflicts.
Validators also attempt concurrent execution; however, without additional information, they may arrive at a
different final state if conflicting transactions execute in different orders. To address this, miners produce
a Block Graph (BG) capturing conflict relations among transactions. Validators use this BG to deterministically
re-execute the transactions and arrive at the same final state.
We implement the framework using Basic Timestamp Ordering (BTO) and Multi-Version Timestamp Ordering (MVTO)
optimistic STMs. BTO and MVTO miners achieve 3.6× and 3.7× average speedups over sequential execution, while
validators achieve 40.8× and 47.1× speedups respectively.
Anjana Parwat Singh,
Priyanka Badiwal,
Rajeev Wankar,
Swaroop Kallakuri,
and C. Raghavendra Rao,
Cloud Service Provider Evaluation System using Fuzzy Rough Set Technique,
In 13th IEEE International Conference on Service-Oriented System Engineering
(SOSE 2019), San Francisco, California, April 2019.
@INPROCEEDINGS{Anjana2019soseEFRSCB,
author = {Anjana, Parwat Singh and Badiwal, Priyanka and Wankar, Rajeev and Kallakuri, Swaroop and Rao, C. Raghavendra},
booktitle = {2019 IEEE International Conference on Service-Oriented System Engineering (SOSE)},
title = {Cloud Service Provider Evaluation System Using Fuzzy Rough Set Technique},
year = {2019},
pages = {187--18709},
doi = {10.1109/SOSE.2019.00033}
}
Cloud Service Providers (CSPs) offer a large variety of scalable, flexible, and cost-efficient services.
However, the diversity of offerings makes it challenging for users to determine and select the most appropriate service.
Users may also need to combine services from multiple providers, leading to management issues involving accounts, security,
support, interfaces, and Service Level Agreements (SLAs).
A Cloud Service Broker (CSB) aware of service capabilities and user Quality of Service (QoS) requirements can significantly
ease this complexity for both users and providers. This work proposes a cloud service brokerage architecture based on fuzzy
rough set techniques for ranking and selecting services according to user QoS requirements.
The system uses fuzzy rough sets for dimensionality reduction and weighted Euclidean distance to rank CSPs. User-assigned
weights prioritize QoS needs, while system-assigned weights emphasize the relative importance of QoS attributes.
Comparative analysis demonstrates that the proposed ranking technique outperforms an existing response-time-based method.
A case study further shows that the proposed approach is scalable, resilient, and achieves better results with reduced
search time.
Anjana Parwat Singh,
Rajeev Wankar,
and C. Raghavendra Rao,
Design of a Cloud Brokerage Architecture Using Fuzzy Rough Set Technique,
In 11th Multi-disciplinary International Workshop on Artificial Intelligence
(MIWAI 2017), Tungku, Brunei, November 2017.
@InProceedings{Anjana2017miwaiFRSCB,
author = {Anjana, Parwat Singh and Wankar, Rajeev and Rao, C. Raghavendra},
title = {Design of a Cloud Brokerage Architecture Using Fuzzy Rough Set Technique},
booktitle = {Multi-disciplinary Trends in Artificial Intelligence},
year = {2017},
publisher = {Springer International Publishing},
address = {Cham},
pages = {54--68},
isbn = {978-3-319-69456-6}
}
Cloud computing offers numerous services to consumers, including infrastructure, platforms, and software.
Due to the diversity of available services, users face significant challenges in ranking and selecting appropriate cloud providers.
Existing solutions using Rough Set Theory (RST) cannot effectively handle numerical values.
A more suitable approach uses Fuzzy Rough Set Theory (FRST), yet no complete Fuzzy-Rough-Set-based brokerage architecture
existed prior to this work. We propose a Fuzzy Rough Set based Cloud Brokerage (FRSCB) architecture for service selection
based on user Quality of Service (QoS) requirements.
The proposed system uses FRST for attribute minimization and search space reduction. QoS attributes are categorized
into functional and non-functional types and classified as static or dynamic. An algorithm is introduced to recommend
potential cloud services to users based on these refined attributes.
The proposed architecture enables efficient, intelligent, and user-aware cloud service selection.
Posters/Short Papers
Parwat Singh Anjana,
Adithya Rajesh Chandrassery,
and Sathya Peri,
An Efficient Approach to Move Elements in a Distributed Geo-Replicated Tree,
In 22nd IEEE/ACM International Symposium on Cluster, Cloud and Internet Computing
(CCGrid 2022), Short Paper, pages 767–770.
@InProceedings{Anjana2022CCGridMoveTreeShort,
author = {Parwat Singh Anjana and Adithya Rajesh Chandrassery and Sathya Peri},
title = {An Efficient Approach to Move Elements in a Distributed Geo-Replicated Tree},
booktitle = {2022 22nd IEEE/ACM International Symposium on Cluster, Cloud and Internet Computing (CCGrid)},
pages = {767--770},
publisher = {IEEE},
year = {2022},
doi = {10.1109/CCGRID54584.2022.00089}
}
The CCGrid 2022 short paper provides a concise summary of the geo-replicated move-tree algorithm,
emphasizing its experimental evaluation as a poster contribution. The work highlights the efficiency
of the proposed move operation and demonstrates its advantages for replicated tree workloads in
distributed environments.
Anjana Parwat Singh,
Efficient Parallel Execution of Block Transactions in Blockchain,
Proceedings of the 22nd International Middleware Conference: Doctoral Symposium
(Middleware 2021), Québec City, Canada, December 2021.
@inproceedings{AnjanaParBlockMiddleware2021,
author = {Anjana, Parwat Singh},
title = {Efficient Parallel Execution of Block Transactions in Blockchain},
year = {2021},
isbn = {9781450391559},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
url = {https://doi.org/10.1145/3491087.3493676},
doi = {10.1145/3491087.3493676},
booktitle = {Proceedings of the 22nd International Middleware Conference: Doctoral Symposium},
pages = {8--11},
numpages = {4},
location = {Québec City, Canada},
series = {Middleware '21}
}
Miners and validators in current blockchains execute block transactions sequentially, which fails to
efficiently utilize modern multi-core architectures, thereby limiting throughput. This work proposes
three approaches to improve blockchain performance by introducing parallel execution of block transactions.
First, we present DiPETrans, a static-analysis-based approach that groups transactions into independent
shards and executes them in parallel using a distributed leader–follower strategy. DiPETrans is evaluated
on 5 million real Ethereum transactions.
Since static analysis cannot precisely identify conflicts, we introduce OptSmart, which exploits
multi-threading on a single multi-core system. Miners execute smart contract transactions (SCTs) concurrently
using optimistic read–write transactional memory systems (RWSTMs). Non-conflicting SCTs are placed in a
concurrent bin, while conflicting ones are stored in a block graph (BG). Validators then deterministically
re-execute SCTs using this BG.
Finally, optimistic object-based STMs (OSTMs) provide even higher concurrency than RWSTMs. We propose the
ObjSC approach using OSTMs, along with a counter-based smart multi-threaded validator (SMV) that can detect
and reject malicious blocks proposed by faulty miners.
Experimental results show that these approaches yield substantial performance improvements over existing
methods.
Anjana Parwat Singh,
Hagit Attiya,
Sweta Kumari,
Sathya Peri,
and Archit Somani,
Efficient Concurrent Execution of Smart Contracts in Blockchains using Object-based Transactional Memory,
A Short Paper for Presentation in the 15th Academic Research and Careers for Students Symposium
(ARCS 2021), Coimbatore, India, February 2021.
Popular blockchains such as Ethereum execute complex transactions through user-defined scripts.
A block typically contains multiple smart contract transactions (SCTs) executed by a miner, and validators
later re-execute these SCTs to determine the block’s correctness during consensus. Miners are incentivized
when valid blocks are successfully appended.
Sequential execution of SCTs fails to exploit multicore processors, reducing throughput. Introducing
concurrency can significantly improve performance. Recent work used Read-Write Software Transactional Memory
Systems (RWSTMs) for concurrent SCT execution. However, Object-based STMs (OSTMs), which use higher-level
abstractions such as hash tables or lists, provide better throughput than RWSTMs.
This paper proposes an efficient concurrency framework based on optimistic single-version OSTMs (SVOSTMs).
A multithreaded miner constructs a Block Graph (BG), capturing object-level conflicts among SCTs, and embeds
it into the block. Validators then re-execute SCTs deterministically in parallel using the BG.
To detect malicious miners (e.g., misreporting BGs), we introduce a Smart Multi-threaded Validator (SMV).
Experimental analysis shows significant performance gains for both miners and validators compared to
state-of-the-art SCT execution frameworks.
Prashansa Agrawal,
Anjana Parwat Singh,
and Sathya Peri,
DeHiDe: Deep Learning-based Hybrid Model to Detect Fake News using Blockchain,
In 22nd International Conference on Distributed Computing and Networking
(ICDCN 2021), Nara, Japan, January 2021.
@inproceedings{Agrawal2021icdcnDeHiDe,
author = {Agrawal, Prashansa and Anjana, Parwat Singh and Peri, Sathya},
title = {DeHiDe: Deep Learning-Based Hybrid Model to Detect Fake News Using Blockchain},
year = {2021},
isbn = {9781450389334},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
url = {https://doi.org/10.1145/3427796.3430003},
doi = {10.1145/3427796.3430003},
booktitle = {International Conference on Distributed Computing and Networking 2021},
pages = {245--246},
numpages = {2},
location = {Nara, Japan},
series = {ICDCN '21}
}
The spread of deep-fakes, misleading information, and fabricated news across social media
has raised significant concerns regarding public influence and information integrity.
Traditional centralized approaches struggle to combat fake news effectively due to lack of
transparency and resistance to manipulation.
This paper introduces DeHiDe, a deep learning–based hybrid model integrated with
blockchain to detect and mitigate fake news. DeHiDe leverages blockchain’s immutability
and decentralized verification to provide a trustworthy framework for legitimate news
dissemination, while the deep learning component enhances robustness and accuracy in
distinguishing real from fake content. The combined approach improves reliability, reduces
vulnerability to tampering, and provides a scalable solution to one of the most critical
challenges in modern digital communication.
Anjana Parwat Singh,
Sweta Kumari, Sathya Peri, Sachin Rathor, and Archit Somani,
Entitling Concurrency to Smart Contracts using Optimistic Transactional Memory,
In 20th International Conference on Distributed Computing and Networking
(ICDCN 2019), Bangalore, India, January 2019.
Recipient of ICDCN 2019 Best Poster Award.
@inproceedings{Anjana2019icdcnEff,
author = {Anjana, Parwat Singh and Kumari, Sweta and Peri, Sathya and Rathor, Sachin and Somani, Archit},
title = {Entitling Concurrency to Smart Contracts Using Optimistic Transactional Memory},
year = {2019},
isbn = {9781450360944},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
url = {https://doi.org/10.1145/3288599.3299723},
doi = {10.1145/3288599.3299723},
booktitle = {Proceedings of the 20th International Conference on Distributed Computing and Networking},
pages = {508},
numpages = {1},
location = {Bangalore, India},
series = {ICDCN '19}
}
Blockchain platforms such as Ethereum execute complex smart contract transactions inside each block.
Miners process these transactions sequentially when proposing a new block, and validators re-execute
them sequentially during verification. This strict serial execution prevents blockchain systems from
utilizing modern multi-core machines efficiently, leading to limited throughput.
This poster explores introducing concurrency into smart contract execution using optimistic
transactional memory (STM). We demonstrate that STM-based speculative parallel execution can
maintain correctness while significantly improving throughput. Our study highlights the feasibility
of optimistic parallelism in blockchain transaction processing, laying the foundation for higher-level
parallel execution frameworks.
Unpublished Manuscripts
Parwat Singh Anjana, Srivatsan Ravi,
Empirical Analysis of Transaction Conflicts in Ethereum and Solana for Parallel Execution,
arXiv preprint arXiv:2505.05358, cs.DC (Distributed, Parallel, and Cluster Computing), 2025.
@article{Anjana2025EmpiricalConflicts,
author = {Anjana, Parwat Singh and Ravi, Srivatsan},
title = {Empirical Analysis of Transaction Conflicts in Ethereum and Solana for Parallel Execution},
journal = {arXiv preprint arXiv:2505.05358},
year = {2025},
url = {https://arxiv.org/abs/2505.05358}
}
This paper presents a comprehensive analysis of historical data across two popular blockchain networks: Ethereum and Solana. Our study focuses on two key aspects: transaction conflicts and the maximum theoretical parallelism within historical blocks. We aim to quantify the degree of transaction parallelism and assess how effectively it can be exploited by systematically examining block-level characteristics, both within individual blocks and across different historical periods. In particular, this study is the first of its kind to leverage historical transactional workloads to evaluate transactional conflict patterns. By offering a structured approach to analyzing these conflicts, our research provides valuable insights and an empirical basis for developing more efficient parallel execution techniques for smart contracts in the Ethereum and Solana virtual machines. Our empirical analysis reveals that historical Ethereum blocks frequently achieve high independence, over 50% in more than 50% of blocks, while Solana historical blocks contain longer conflict chains, comprising ∼59% of the block size compared to ∼18% in Ethereum, reflecting fundamentally different parallel execution dynamics.
Priyanka Badiwal,
Anjana Parwat Singh,
Rajeev Wankar,
and C. Raghavendra Rao,
DRONA: A Data-driven Randomized Algorithm for Complex Optimization Problems.
Recent Projects
Parallel Execution
SupraAdaptive – a high-throughput workload adaptive parallel execution engine powering
Supra's Multi-VM (EVM, MoveVM and SVM) [Coming Soon].
Historical Conflict Analysis of Blockchain Transactions, large-scale empirical study of Ethereum
and Solana read–write conflicts, parallelism potential, and execution bottlenecks.
[
arXiv:2505.05358
]
I am very fortunate to collaborate with outstanding researchers, mentors, and colleagues whose guidance,
insights, and discussions have significantly shaped my work in distributed systems, transactional memory,
and high-performance blockchain execution.
Awarded Institute Postdoctoral Fellowship for 6 months (February 2022 to July 2022) at
IIT Hyderabad.
Won Best Poster Award at the Doctoral Symposium (ICDCN 2019) for the poster
Entitling Concurrency to Smart Contracts Using Optimistic Transactional Memory.
[
ICDCN 2019
]
Qualified UGC NET (India) – December 2015 for Lectureship.
[
Score Card
]
Qualified GATE (India) for four consecutive years (2014–2017),
and awarded the GATE Scholarship for pursuing M.Tech (2014–2016).
