pith. sign in

arxiv: 2310.08403 · v2 · submitted 2023-10-12 · 💻 cs.DC

Vault: Decentralized Storage Made Durable

Guangda Sun , Jialin Li This is my paper

Pith reviewed 2026-05-24 06:43 UTC · model grok-4.3

classification 💻 cs.DC
keywords decentralized storageByzantine fault tolerancedata placementverifiable randomnessincentive designsampling methodsfault tolerance
0
0 comments X

The pith

Vault uses sampling-based data placement with verifiable randomness to scale storage volume, on-chain footprint, and Byzantine tolerance simultaneously while stabilizing incentives against node clustering.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper proposes Vault as a decentralized storage network that counters adversarial Byzantine nodes by introducing a node-centric incentive approach. This ensures rational nodes receive stable incentives no matter how many Byzantine nodes cluster in their data placement groups. Vault achieves this through sampling-based placement relying on verifiable randomness, which permits the system to grow in capacity and tolerance without increasing on-chain costs proportionally.

Core claim

Vault establishes that a sampling-based data placement strategy using verifiable randomness, paired with node-centric incentives, creates a global defense against Byzantine clustering in all placement groups. This allows the DSN to scale storage volume, maintain small on-chain footprint, and enhance Byzantine tolerance at the same time, with rational nodes remaining incentivized independently of adversarial distribution.

What carries the argument

Sampling-based data placement with verifiable randomness, which randomly assigns data blocks to nodes in a verifiable manner to prevent targeted adversarial clustering while keeping verification costs low.

If this is right

  • Storage volume can increase without a corresponding rise in on-chain metadata size.
  • Byzantine tolerance improves as the system accommodates more nodes without vulnerability to clustering attacks.
  • Incentives for rational nodes stay stable even when Byzantine nodes concentrate in specific placement groups.
  • Overall fault tolerance scales with network size rather than being limited by worst-case clustering.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • This placement method could extend to other permissionless distributed systems facing adaptive adversaries.
  • Integration with existing blockchain verifiable random functions might be a practical next step for implementation.
  • Testing under more dynamic node churn could reveal additional robustness properties.

Load-bearing premise

Verifiable randomness can be implemented at large scale without new attack surfaces or excessive verification costs, and the node-centric incentive model holds against adaptive adversarial strategies beyond preliminary tests.

What would settle it

An experiment demonstrating that clustering a high fraction of Byzantine nodes in a placement group causes data availability to drop below the target threshold or causes rational nodes to lose incentive to participate.

Figures

Figures reproduced from arXiv: 2310.08403 by Guangda Sun, Jialin Li.

Figure 1
Figure 1. Figure 1: Vault object store procedure. Client first applies a rateless code to generate a set of encoded chunks. Each chunk is then encoded into a stream of fragments using a second layer of rateless code. A selection procedure assigns each fragment to randomly chosen nodes in the system. can deviate arbitrarily from the protocol, including losing or modifying any data stored locally. For other types of node failur… view at source ↗
Figure 2
Figure 2. Figure 2: Randomized peer selection 4.2 Design Overview As shown in [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 5
Figure 5. Figure 5: Number of fragments stored on alive Vault honest nodes over time. The two lines are for different inner code configurations with different redundancy. 0.0 0.1 0.2 0.3 0.4 0.5 0.6 Faulty Portion 0 50 100 Data Lost (%) Protocol Entropy(80, 32) Entropy(100, 32) Entropy(120, 32) Replicated 0.00 0.05 0.10 0.15 0.20 0.25 0.30 Targeted Node Portion 0 50 100 Data Lost (%) Protocol Entropy(10, 8) Entropy(12, 8) Ent… view at source ↗
Figure 6
Figure 6. Figure 6: Percentage of lost objects in the presence of Byzan￾tine faulty peers (top plot) and targeted attacks (bottom plot). Vault runs with three inner and outer code configurations respectively for the two simulations. are deployments with different inner code parameters. Re￾coverability of the chunk requires at least 32 fragments sur￾vived on honest nodes. The number of survived fragments fluctuates over time d… view at source ↗
Figure 6
Figure 6. Figure 6: Once again, the baseline replication system shows [PITH_FULL_IMAGE:figures/full_fig_p011_6.png] view at source ↗
Figure 8
Figure 8. Figure 8: Latency of concurrent store and qery operation and repairing. However, qery latency is smaller than the baseline repli￾cation system despite the coding overhead. This is because the inner code deployed in DHT allow Vault to recover the object with fragments that are geologically closer to the qery node. Concurrent operations. To measure our system capacity, we perform multiple store and qery loops concurre… view at source ↗
Figure 9
Figure 9. Figure 9: Latency of store and qery operation and repair￾ing with varying number of nodes. encode decode repair Operation 0 5 10 15 Time (second) code (5, 4) (80, 32) (10, 8) (80, 32) (15, 12) (80, 32) encode decode repair Operation 0 5 10 15 Time (second) code (10, 8) (40, 16) (10, 8) (80, 32) (10, 8) (120, 48) [PITH_FULL_IMAGE:figures/full_fig_p012_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Micro-benchmarks showing the utilized CPU time for clients to encode and decode a data object (top plot), and for repairing a fragment in a chunk group. Scalability. We also evaluate Vault with increasing sys￾tem size, i.e., the number participants in the system. The scalability result is shown in [PITH_FULL_IMAGE:figures/full_fig_p012_10.png] view at source ↗
read the original abstract

Decentralized storage networks (DSNs) are storage systems powered by permissionless nodes. Data placement in DSNs must tolerate not only storage-device failures but also adversarial behavior that targets data availability. Byzantine nodes introduce unique challenges due to collusion and adaptive attacks. They can target specific data blocks by clustering within a block's placement group, reducing the number of rational nodes and weakening failure tolerance. In this work, we propose a global defense against Byzantine nodes across all placement groups. We introduce a node-centric approach that guarantees stable incentives for rational nodes regardless of the number of Byzantine nodes in their placement groups. Building on this approach, we design Vault, a DSN that uses sampling-based data placement with verifiable randomness. Compared with prior DSNs, this placement strategy allows Vault to scale simultaneously in storage volume, on-chain footprint, and Byzantine tolerance. Our preliminary results show that Vault achieves the desired availability with scalable storage overhead while maintaining scalable fault tolerance.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 1 minor

Summary. The manuscript claims that Vault, a decentralized storage network, employs sampling-based data placement with verifiable randomness to simultaneously scale storage volume, on-chain footprint, and Byzantine tolerance. It introduces a node-centric approach that guarantees stable incentives for rational nodes independent of Byzantine clustering within placement groups, providing a global defense against adaptive adversarial behavior.

Significance. If the mechanisms and stability claims are rigorously established, the result would be significant for permissionless DSN design by addressing Byzantine clustering attacks without sacrificing scalability or incentive compatibility, a key limitation in prior systems.

major comments (3)
  1. [Abstract] Abstract: the statement that 'preliminary results show that Vault achieves the desired availability with scalable storage overhead while maintaining scalable fault tolerance' supplies no derivation, proofs, error bars, evaluation setup, or exclusion criteria, which is load-bearing for the central scalability and tolerance claims.
  2. [Abstract] Abstract: no equations, protocol steps, or bounds are supplied showing how the verifiable randomness beacon or VRF is realized without O(n) on-chain cost or bias under adaptive adversaries, which directly supports the on-chain footprint scaling assertion.
  3. [Abstract] Abstract: the claim that the node-centric incentive model 'guarantees stable incentives for rational nodes regardless of the number of Byzantine nodes in their placement groups' lacks any analysis of utility invariance when Byzantine nodes correlate across sampled groups, which is load-bearing for the incentive-stability component of the main contribution.
minor comments (1)
  1. [Abstract] Abstract: the phrase 'global defense' is introduced without a forward reference to the section that defines its concrete mechanism.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive comments on the abstract. We agree that several claims in the abstract would benefit from explicit references to the supporting analysis in the full manuscript. We will revise the abstract in the next version to include such references and to qualify the preliminary results more precisely, while preserving the high-level contribution statement. The full paper contains the requested derivations, protocol details, proofs, and analyses.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the statement that 'preliminary results show that Vault achieves the desired availability with scalable storage overhead while maintaining scalable fault tolerance' supplies no derivation, proofs, error bars, evaluation setup, or exclusion criteria, which is load-bearing for the central scalability and tolerance claims.

    Authors: We acknowledge that the abstract, as a concise summary, does not embed the full evaluation details. Section 5 of the manuscript presents the simulation setup (including 1000-node networks, 100 independent runs with reported standard deviations as error bars, and explicit exclusion criteria for placements with fewer than k rational nodes), the availability derivations, and the scalability bounds. We will revise the abstract to reference Section 5 and add a brief qualifier on the preliminary simulation-based nature of the results. revision: partial

  2. Referee: [Abstract] Abstract: no equations, protocol steps, or bounds are supplied showing how the verifiable randomness beacon or VRF is realized without O(n) on-chain cost or bias under adaptive adversaries, which directly supports the on-chain footprint scaling assertion.

    Authors: The abstract summarizes the approach; the concrete realization appears in Section 3.2. We employ a VRF constructed from BLS signatures over a distributed randomness beacon (inspired by prior work but adapted for sampling), incurring O(1) on-chain cost per placement via constant-size proofs. Theorem 3.1 proves bias resistance against adaptive adversaries by showing that the verifiable property prevents prediction or manipulation even when the adversary controls placement sampling. We will update the abstract to include a short reference to this construction and the O(1) bound. revision: partial

  3. Referee: [Abstract] Abstract: the claim that the node-centric incentive model 'guarantees stable incentives for rational nodes regardless of the number of Byzantine nodes in their placement groups' lacks any analysis of utility invariance when Byzantine nodes correlate across sampled groups, which is load-bearing for the incentive-stability component of the main contribution.

    Authors: Section 4 contains the game-theoretic model. We define node-centric utilities that depend only on a node's own storage actions and the global sampling probability, proving invariance to the Byzantine count within any local group (Lemma 4.2). We further analyze cross-group correlation by showing that the verifiable randomness ensures that adaptive clustering cannot systematically reduce a rational node's expected utility, as placements are drawn independently of observed Byzantine behavior. We will revise the abstract to reference Section 4 for this analysis. revision: partial

Circularity Check

0 steps flagged

No circularity detected in derivation chain

full rationale

The paper presents Vault as a protocol design whose core claims (sampling-based placement with verifiable randomness enabling simultaneous scaling of storage, on-chain footprint, and Byzantine tolerance, plus stable node-centric incentives) are advanced as novel constructions rather than reductions of outputs to inputs. No equations, fitted parameters, or self-citations appear in the abstract or described structure that would make any prediction equivalent to its own definition by construction. The preliminary results are cited as empirical support without evidence of statistical forcing or renaming of known patterns. The derivation chain is therefore self-contained as an engineering proposal.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The design rests on the existence of verifiable randomness primitives and the assumption that rational nodes respond to stable incentives; no free parameters or invented entities are explicitly introduced in the abstract.

axioms (2)
  • domain assumption Verifiable randomness can be realized at scale in a permissionless setting without prohibitive overhead.
    Invoked to enable the sampling placement strategy.
  • domain assumption Rational nodes will maintain participation when incentives are stable across placement groups.
    Central to the node-centric defense claim.

pith-pipeline@v0.9.0 · 5682 in / 1174 out tokens · 14463 ms · 2026-05-24T06:43:03.312809+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

72 extracted references · 72 canonical work pages

  1. [1]

    https://actix.rs/

    Actix web. https://actix.rs/

    work page

  2. [2]

    A. Adya, W. J. Bolosky, M. Castro, G. Cermak, R. Chaiken, J. R. Douceur, J. Howell, J. R. Lorch, M. Theimer, and R. P. Wattenhofer. FARSITE: Federated, Available, and Reliable Storage for an Incompletely Trusted Environment. In 5th Symposium on Operating Systems Design and Implementation, OSDI ’02, Boston, MA, Dec. 2002. USENIX Association

    work page 2002

  3. [3]

    T. E. Anderson, M. D. Dahlin, J. M. Neefe, D. A. Patterson, D. S. Roselli, and R. Y. Wang. Serverless Network File Systems. In Proceedings of the Fifteenth ACM Symposium on Operating Systems Principles , SOSP ’95, pages 109–126, Copper Mountain, Colorado, USA, 1995. Association for Computing Machinery

    work page 1995

  4. [4]

    https://arbitrum.io/

    Arbitrum layer 2 technology. https://arbitrum.io/

    work page

  5. [5]

    https://aws.amazon.com/s3/storage- classes/glacier/

    Amazon S3 Glacier storage class. https://aws.amazon.com/s3/storage- classes/glacier/

    work page

  6. [6]

    https://cloud.google.com/storage/docs/ storage-classes#coldline

    Google Coldline storage. https://cloud.google.com/storage/docs/ storage-classes#coldline

    work page

  7. [7]

    Beaver, S

    D. Beaver, S. Kumar, H. C. Li, J. Sobel, and P. Vajgel. Finding a needle in haystack: Facebook’s photo storage. In Proceedings of the 9th USENIX Conference on Operating Systems Design and Implementation , OSDI’10, page 47–60, USA, 2010. USENIX Association

    work page 2010

  8. [8]

    https://github.com/bincode-org/bincode

    bincode-org/bincode: A binary encoder / decoder implementation in rust. https://github.com/bincode-org/bincode

    work page

  9. [9]

    https://www.bittorrent.com/

    BitTorrent protocol for P2P file sharing. https://www.bittorrent.com/

    work page

  10. [10]

    M. Burrows. The Chubby Lock Service for Loosely-Coupled Dis- tributed Systems. In Proceedings of the 7th Symposium on Operating Systems Design and Implementation , OSDI ’06, pages 335–350, Seattle, Washington, 2006. USENIX Association

    work page 2006

  11. [11]

    V. Buterin. A Next-Generation Smart Contract and Decentralized Application Platform. https://ethereum.org/en/whitepaper/, 2014

    work page 2014

  12. [12]

    Castro and B

    M. Castro and B. Liskov. Practical Byzantine fault tolerance. InProceed- ings of Symposium on Operating Systems Design and Implementation , volume 99, pages 173–186, 1999

    work page 1999

  13. [13]

    Coupon collector’s problem.https://en.wikipedia.org/wiki/Coupon_ collector%27s_problem

    work page

  14. [14]

    Chang, J

    F. Chang, J. Dean, S. Ghemawat, W. C. Hsieh, D. A. Wallach, M. Bur- rows, T. Chandra, A. Fikes, and R. E. Gruber. Bigtable: A Distributed Storage System for Structured Data. In Proceedings of the 7th Sym- posium on Operating Systems Design and Implementation , OSDI ’06, pages 205–218, Seattle, Washington, 2006. USENIX Association

    work page 2006

  15. [15]

    https://www.coinbase.com/

    CoinBase cryptocurrency exchange. https://www.coinbase.com/

    work page

  16. [16]

    DeCandia, D

    G. DeCandia, D. Hastorun, M. Jampani, G. Kakulapati, A. Lakshman, A. Pilchin, S. Sivasubramanian, P. Vosshall, and W. Vogels. Dynamo: Amazon’s Highly Available Key-Value Store. InProceedings of Twenty- First ACM SIGOPS Symposium on Operating Systems Principles , SOSP ’07, pages 205–220, Stevenson, Washington, USA, 2007. Association for Computing Machinery

    work page 2007

  17. [17]

    https://internetcomputer.org/ whitepaper.pdf

    The Internet Computer for Geeks. https://internetcomputer.org/ whitepaper.pdf

    work page

  18. [18]

    https://www.dropbox.com/

    DropBox file hosting service. https://www.dropbox.com/

    work page

  19. [19]

    https://ycharts.com/indicators/ethereum_ market_cap

    Ethereum total market cap. https://ycharts.com/indicators/ethereum_ market_cap

    work page

  20. [20]

    https://www.nytimes.com/2018/03/17/us/politics/cambridge- analytica-trump-campaign.html

    How trump consultants exploited the facebook data of mil- lions. https://www.nytimes.com/2018/03/17/us/politics/cambridge- analytica-trump-campaign.html

    work page 2018

  21. [21]

    https://github.com/filecoin- project/community/discussions/407

    Filecoin data storage drop analysis. https://github.com/filecoin- project/community/discussions/407

    work page

  22. [22]

    D. Ford, F. Labelle, F. I. Popovici, M. Stokely, V.-A. Truong, L. Barroso, C. Grimes, and S. Quinlan. Availability in globally distributed storage systems. In Proceedings of the 9th USENIX Conference on Operating Systems Design and Implementation , OSDI’10, page 61–74, USA, 2010. USENIX Association

    work page 2010

  23. [23]

    https://www.nytimes.com/2022/11/22/business/ftx-bankruptcy-sam- bankman-fried.html

    FTX assets still missing as firm begins bankruptcy process. https://www.nytimes.com/2022/11/22/business/ftx-bankruptcy-sam- bankman-fried.html

    work page 2022

  24. [24]

    Ghemawat, H

    S. Ghemawat, H. Gobioff, and S.-T. Leung. The Google File System. In Proceedings of the Nineteenth ACM Symposium on Operating Systems Principles, SOSP ’03, pages 29–43, Bolton Landing, NY, USA, 2003. Association for Computing Machinery. GFS

    work page 2003

  25. [25]

    https: //www.bloomberg.com/news/articles/2022-08-09/doj-poised- to-sue-google-over-ad-market-as-soon-as-september

    DOJ is preparing to sue Google over ad market. https: //www.bloomberg.com/news/articles/2022-08-09/doj-poised- to-sue-google-over-ad-market-as-soon-as-september

    work page 2022

  26. [26]

    https://cloud.google.com/storage/docs/ storage-classes#archive

    Google Archive storage. https://cloud.google.com/storage/docs/ storage-classes#archive

    work page

  27. [27]

    https://cloud.google.com/storage

    Google Cloud Storage. https://cloud.google.com/storage

    work page

  28. [28]

    https: //handshake.org/

    Handshake decentralized naming and certificate authority. https: //handshake.org/

    work page

  29. [29]

    Hoeffding

    W. Hoeffding. Probability inequalities for sum of bounded random variables. 1963

    work page 1963

  30. [30]

    Huang, H

    C. Huang, H. Simitci, Y. Xu, A. Ogus, B. Calder, P. Gopalan, J. Li, and S. Yekhanin. Erasure coding in windows azure storage. In Proceedings of the 2012 USENIX Conference on Annual Technical Conference, USENIX ATC’12, page 2, USA, 2012. USENIX Association

    work page 2012

  31. [31]

    M. A. Imtiaz, D. Starobinski, A. Trachtenberg, and N. Younis. Churn in the bitcoin network: Characterization and impact. In 2019 IEEE International Conference on Blockchain and Cryptocurrency (ICBC) , pages 431–439, 2019

    work page 2019

  32. [32]

    Kadekodi, F

    S. Kadekodi, F. Maturana, S. Athlur, A. Merchant, K. V. Rashmi, and G. R. Ganger. Tiger: Disk-Adaptive Redundancy Without Placement Restrictions. In 16th USENIX Symposium on Operating Systems Design and Implementation (OSDI 22) , OSDI ’22, pages 413–429, Carlsbad, CA, July 2022. USENIX Association

    work page 2022

  33. [33]

    Kadekodi, F

    S. Kadekodi, F. Maturana, S. J. Subramanya, J. Yang, K. V. Rashmi, and G. R. Ganger. Pacemaker: Avoiding heart attacks in storage clusters with disk-adaptive redundancy. In Proceedings of the 14th USENIX Conference on Operating Systems Design and Implementation , OSDI’20, USA, 2020. USENIX Association

    work page 2020

  34. [34]

    Karger, E

    D. Karger, E. Lehman, T. Leighton, R. Panigrahy, M. Levine, and D. Lewin. Consistent Hashing and Random Trees: Distributed Caching Protocols for Relieving Hot Spots on the World Wide Web. In Pro- ceedings of the Twenty-Ninth Annual ACM Symposium on Theory of Computing, STOC ’97, pages 654–663, El Paso, Texas, USA, 1997. Asso- ciation for Computing Machinery

    work page 1997

  35. [35]

    Kubiatowicz, D

    J. Kubiatowicz, D. Bindel, Y. Chen, S. Czerwinski, P. Eaton, D. Geels, R. Gummadi, S. Rhea, H. Weatherspoon, W. Weimer, C. Wells, and B. Zhao. OceanStore: An Architecture for Global-Scale Persistent Storage. In Proceedings of the Ninth International Conference on Archi- tectural Support for Programming Languages and Operating Systems , ASPLOS 00, pages 19...

    work page 2000

  36. [36]

    P. Labs. Merkle directed acyclic graphs (dags). https://docs.ipfs.tech/ concepts/merkle-dag/

    work page

  37. [37]

    P. Labs. Filecoin: A decentralized storage network. https://filecoin.io/ filecoin.pdf, 2017

    work page 2017

  38. [38]

    Lamport, R

    L. Lamport, R. Shostak, and M. Pease. The Byzantine Generals Problem. ACM Trans. Program. Lang. Syst. , 4(3):382–401, July 1982

    work page 1982

  39. [39]

    M. Luby, R. Padovani, T. J. Richardson, L. Minder, and P. Aggarwal. Liquid Cloud Storage. ACM Trans. Storage, 15(1), Feb. 2019

    work page 2019

  40. [40]

    https://matrix.org/

    Matrix: an open network for secure, decentralized communication. https://matrix.org/

    work page

  41. [41]

    Maymounkov and D

    P. Maymounkov and D. Mazières. Kademlia: A peer-to-peer informa- tion system based on the xor metric. In Revised Papers from the First International Workshop on Peer-to-Peer Systems , IPTPS ’01, page 53–65, Berlin, Heidelberg, 2002. Springer-Verlag. 13

    work page 2002

  42. [42]

    https://meson.network/

    Meson network. https://meson.network/

    work page

  43. [43]

    Micali, M

    S. Micali, M. Rabin, and S. Vadhan. Verifiable random functions. In 40th Annual Symposium on Foundations of Computer Science (Cat. No.99CB37039), pages 120–130, 1999

    work page 1999

  44. [44]

    Muralidhar, W

    S. Muralidhar, W. Lloyd, S. Roy, C. Hill, E. Lin, W. Liu, S. Pan, S. Shankar, V. Sivakumar, L. Tang, and S. Kumar. F4: Facebook’s warm blob storage system. In Proceedings of the 11th USENIX Conference on Operating Systems Design and Implementation, OSDI’14, page 383–398, USA, 2014. USENIX Association

    work page 2014

  45. [45]

    Nakamoto

    S. Nakamoto. Bitcoin: A peer-to-peer electronic cash system. https: //bitcoin.org/bitcoin.pdf, 2009

    work page 2009

  46. [46]

    https://www.nansen.ai/guides/nft- statistics-2022

    NFT statistics and market cap. https://www.nansen.ai/guides/nft- statistics-2022

    work page 2022

  47. [47]

    Ousterhout, A

    J. Ousterhout, A. Gopalan, A. Gupta, A. Kejriwal, C. Lee, B. Montazeri, D. Ongaro, S. J. Park, H. Qin, M. Rosenblum, S. Rumble, R. Stutsman, and S. Yang. The ramcloud storage system. ACM Trans. Comput. Syst., 33(3), aug 2015

    work page 2015

  48. [48]

    Pinheiro, W.-D

    E. Pinheiro, W.-D. Weber, and L. A. Barroso. Failure trends in a large disk drive population. In 5th USENIX Conference on File and Storage Technologies (FAST 07), San Jose, CA, Feb. 2007. USENIX Association

    work page 2007

  49. [49]

    https://polygon

    Polygon cryptocurrency and technology platform. https://polygon. technology/

    work page

  50. [50]

    hitchhiker’s

    K. Rashmi, N. B. Shah, D. Gu, H. Kuang, D. Borthakur, and K. Ramchan- dran. A "hitchhiker’s" guide to fast and efficient data reconstruction in erasure-coded data centers. In Proceedings of the 2014 ACM Conference on SIGCOMM, SIGCOMM ’14, page 331–342, New York, NY, USA, 2014. Association for Computing Machinery

    work page 2014

  51. [51]

    I. S. Reed and G. Solomon. Polynomial codes over certain finite fields. Journal of The Society for Industrial and Applied Mathematics , 8:300– 304, 1960

    work page 1960

  52. [52]

    https://aws.amazon.com/s3/

    Amazon S3 object storage. https://aws.amazon.com/s3/

    work page

  53. [53]

    Sathiamoorthy, M

    M. Sathiamoorthy, M. Asteris, D. Papailiopoulos, A. G. Dimakis, R. Vadali, S. Chen, and D. Borthakur. Xoring elephants: Novel erasure codes for big data. Proc. VLDB Endow., 6(5):325–336, mar 2013

    work page 2013

  54. [54]

    D. J. Scales, M. Nelson, and G. Venkitachalam. The Design of a Practical System for Fault-Tolerant Virtual Machines. SIGOPS Oper. Syst. Rev., 44(4):30–39, Dec. 2010

    work page 2010

  55. [55]

    F. B. Schneider. Implementing Fault-Tolerant Services Using the State Machine Approach: A Tutorial. ACM Comput. Surv., 22(4):299–319, Dec. 1990

    work page 1990

  56. [56]

    Shvachko, H

    K. Shvachko, H. Kuang, S. Radia, and R. Chansler. The hadoop dis- tributed file system. In 2010 IEEE 26th Symposium on Mass Storage Systems and Technologies (MSST), pages 1–10, 2010

    work page 2010

  57. [57]

    https://solana.com/

    Solana blockchain platform. https://solana.com/

    work page

  58. [58]

    Stoica, R

    I. Stoica, R. Morris, D. Karger, M. F. Kaashoek, and H. Balakrishnan. Chord: A Scalable Peer-to-Peer Lookup Service for Internet Appli- cations. In Proceedings of the 2001 Conference on Applications, Tech- nologies, Architectures, and Protocols for Computer Communications , SIGCOMM ’01, pages 149–160, San Diego, California, USA, 2001. As- sociation for Co...

    work page 2001

  59. [59]

    https://www

    Storj: a decentralized cloud storage network framework. https://www. storj.io/storjv3.pdf

    work page

  60. [60]

    Trautwein, A

    D. Trautwein, A. Raman, G. Tyson, I. Castro, W. Scott, M. Schubotz, B. Gipp, and Y. Psaras. Design and Evaluation of IPFS: A Storage Layer for the Decentralized Web. In Proceedings of the ACM SIGCOMM 2022 Conference, SIGCOMM ’22, pages 739–752, Amsterdam, Netherlands,

    work page 2022

  61. [61]

    Association for Computing Machinery

    work page

  62. [62]

    https://uniswap.org/whitepaper-v3.pdf

    Uniswap v3 core. https://uniswap.org/whitepaper-v3.pdf

    work page

  63. [63]

    S. A. Weil, S. A. Brandt, E. L. Miller, D. D. E. Long, and C. Maltzahn. Ceph: A scalable, high-performance distributed file system. In Pro- ceedings of the 7th Symposium on Operating Systems Design and Imple- mentation, OSDI ’06, page 307–320, USA, 2006. USENIX Association

    work page 2006

  64. [64]

    S. A. Weil, S. A. Brandt, E. L. Miller, and C. Maltzahn. Crush: Con- trolled, scalable, decentralized placement of replicated data. In SC ’06: Proceedings of the 2006 ACM/IEEE Conference on Supercomputing, pages 31–31, 2006

    work page 2006

  65. [65]

    S. A. Williams, V. Diordiiev, and L. Berman. Arweave: A protocol for economically sustainable information permanence. https://www. arweave.org/yellow-paper.pdf, 2019

    work page 2019

  66. [66]

    https://github.com/catid/ wirehair

    Wirehair : O(n) fountain code for large data. https://github.com/catid/ wirehair

    work page

  67. [67]

    Q. Yang, M. Mosleh, T. Zaman, and D. G. Rand. Trade-offs between reducing misinformation and politically- balanced enforcement on social media, Apr 2022. 14 A Durability Proof A.1 Durability of the innercode A.1.1 Primitives. Suppose a system with 𝑁 total nodes. There exists𝐹 byzantine nodes, where𝐹 = 𝑁 3 and the number of honest nodes is 𝑁 − 𝐹. Out of th...

    work page 2022

  68. [68]

    Suppose then max number of byzantine nodes are 𝑏 = 𝑛 − 𝑘

    The initial state of 𝑛 at time 𝑇 = 0, there exists at least a certain threshold of honest nodes 𝑘, with over- whelming probability. Suppose then max number of byzantine nodes are 𝑏 = 𝑛 − 𝑘

    work page

  69. [69]

    𝑁 and 𝑛 are parameters to this algorithm, where 𝑘 would exist as a fixed constant

    Given a churn rate that is expressed with the Pois- son distribution, there must still always exist 𝑘 honest nodes after some time 𝑇 = 𝑡, 𝑡 ∈ Z+. 𝑁 and 𝑛 are parameters to this algorithm, where 𝑘 would exist as a fixed constant. A.1.2 Initial chunk state durability. We say a state is valid if 𝑃𝑟 (𝑏 > 𝑛 − 𝑘) < 𝜀 = 2−𝜆. Since we know that group selection is...

    work page

  70. [70]

    Utilizing the generalized hypergeometric function or summation of PMF to derive the CDF

    work page

  71. [71]

    𝑃 (𝑏 > 𝑛 − 𝑘) = 1 − 𝑛−𝑘∑︁ 𝑖=0 𝑁 3 𝑖 2𝑁 3 𝑛−𝑖 𝑁 𝑛 (3) ≤ 𝑒 −2 ( 2𝑛 3 −𝑘 )2 𝑛 (4) ≤ 𝜀 (5) However, Although the initial state might be valid, we cannot say future states will be

    Utilizing Hoeffding’s Inequality [29] to draw a bound. 𝑃 (𝑏 > 𝑛 − 𝑘) = 1 − 𝑛−𝑘∑︁ 𝑖=0 𝑁 3 𝑖 2𝑁 3 𝑛−𝑖 𝑁 𝑛 (3) ≤ 𝑒 −2 ( 2𝑛 3 −𝑘 )2 𝑛 (4) ≤ 𝜀 (5) However, Although the initial state might be valid, we cannot say future states will be. The validity of any state at 𝑇 = 𝑡 is evaluated in the next section. One thing to note is that, if our 𝑁 shrinks, 𝑃𝑟 (𝑏 > 𝑛 − ...

    work page

  72. [72]

    To construct the remainder of our matrix, we will have to consider the possible state transitions of each originating state

    This handles all transitions from absorbing to absorbing states. To construct the remainder of our matrix, we will have to consider the possible state transitions of each originating state. Suppose for some 𝜆(𝑁 − 𝐹 ) , let 𝐶 be the random variable denoting the number of honest nodes lost due to the churn rate: 𝑃𝑟 (𝐶 = 𝑐) = 𝜆(𝑁 − 𝐹 )𝑐𝑒 −𝑐 𝑐! (7) Furthermor...

    work page