Measuring Ransomware Lateral Movement Susceptibility via Privilege-Weighted Adjacency Matrix Exponentiation
Satyam Tyagi · 2025
Read Measuring Ransomware Lateral Movement Susceptibility via Privilege-Weighted Adjacency Matrix Exponentiation by Satyam Tyagi (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-21005.txt.
Introduction
Satyam Tyagi and a co-author published this cybersecurity paper in 2025. Ransomware does its worst damage by spreading sideways through a network from one compromised machine to many. The authors build a graph model of that spread: machines are points, and reachable services such as remote desktop form the connecting edges. Each service gets a pivot potential, a score for how easily an attacker can hop onward through it.
Repeatedly applying the model yields a matrix of compromise chances between every pair of machines within a set number of hops. From it come two headline scores: an overall susceptibility rating and an estimate of the blast radius, the share of assets an average attack would reach. On real enterprise snapshots, cutting the highest-risk service links shrank both scores the most. The authors pitch the framework for judging segmentation and ranking defenses.
Catalog record
| CREATOR | Satyam Tyagi |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 33 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-21005
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Measuring Ransomware Lateral Movement Susceptibility via Privilege-Weighted Adjacency Matrix Exponentiation Satyam Tyagi Abstract Abstract: Ransomware impact hinges on how easily an intruder can move laterally and spread to the maximum number of assets. We present a graph-theoretic method to measure lateral-movement susceptibility and estimate blast radius. We build a directed multigraph where vertices represent assets and edges represent reachable services (e.g., RDP/SSH) between them. We model lateral movement as a probabilistic process using a pivot potential factor \pi(s) for each service. This allows us to iteratively compute a K -hop compromise probability matrix that captures how compromise propagates through the network. Metrics derived from this model include: (1) Lateral-Movement Susceptibility (LMS K ): the average probability of a successful lateral movement between any two assets (0-1 scale); and (2) Blast-Radius Estimate (BRE K ): the expected percentage of assets compromised in an average attack scenario. Interactive control (SSH 22, RDP 3389) gets higher \pi(s) than app-only ports (MySQL 3306, MSSQL 1433), which seldom enable pivoting without an RCE. Across anonymized enterprise snapshots, pruning high- \pi(s) edges yields the largest LMS K /BRE K drop, aligning with CISA guidance, MITRE ATT&CK (TA0008: Lateral Movement), and NIST SP 800-207. The framework evaluates (micro)segmentation and helps prioritize controls that reduce lateral movement susceptibility and shrink blast radius. Keywords: adjacency matrix; multi-hop reachability; privilege weighting; lateral movement; microsegmentation; blast radius; ransomware; Zero Trust Mathematics Subject Classification : 05C50; 05C90; 94C15 1. Introduction Ransomware impact is determined less by where an adversary first lands and more by how easily they can move laterally to reach high-value assets and propagate encryption. Classic reachability (Boolean adjacency, k -hop closure) reveals where paths exist, but it can overstate risk by treating an SSH or RDP foothold as equivalent to a low-privilege application socket. In practice, interactive remote services (e.g., RDP/SSH) enable immediate pivoting; app-only ports typically demand additional exploits. Our goal is a probability-weighted alternative that measures the ease of movement —and the expected blast radius—using only topology plus coarse service-class information. We introduce a privilege- and account-aware adjacency model that composes per-service “pivot potentials” \pi(s)\!\in\![0,1] with multiplicative step success and a noisy-OR union across alternative routes. Iterating this algebraic path closure yields (i) a k -hop (or \varepsilon -fixed-point) compromise probability matrix and (ii) two deployment metrics: LMS K (lateral-movement susceptibility; mean off-diagonal) and BRE K (blast-radius estimate; mean including the diagonal). CVE-free lateral movement via valid accounts. Once an adversary holds valid credentials and an interactive remote service is exposed (RDP/SSH), lateral movement does not require unpatched vulnerabilities: adversaries log into remote services and act as the authenticated user [ MITRE 2025f ; MITRE 2025b ; MITRE 2025c ] . Patch status is orthogonal to pivotability when privileged landing is possible; hence our operators weight service exposure and privilege class , not CVE likelihoods. Contributions. • Privilege- and account-aware path closure. We formalize lateral movement over [0,1] with step-wise multiplication and noisy-OR unions, explicitly capturing valid-account pivots on interactive services [ MITRE 2025b ] . • Service-class lifting. We encode exact “land-then-pivot” semantics by lifting states to (\text{node},\text{class}) , preserving first-hop reach while applying class-conditioned pivot probabilities thereafter. • Actionable metrics. We define LMS K and BRE K for policy comparison and segmentation planning. • Theoretical bounds. We situate our iteration within algebraic path theory, giving a series upper bound via U_{K}=A+AW+\cdots+AW^{K-1} and a single-path product lower bound. • Alignment with guidance. Our empirical reductions in LMS K /BRE K under RDP/SSH constraints match Zero Trust and ransomware guidance to limit remote desktop services and constrain lateral movement [ Rose et al. 2020 ; CISA 2025a ; CISA 2025b ; CISA 2025c ] . Relation to Zero Trust and segmentation. Zero Trust (NIST SP 800-207) calls for minimizing implicit trust and continuously evaluating access in context [ Rose et al. 2020 ] . Our metrics quantify the residual ease of movement under candidate policies, complementing microsegmentation by highlighting edges whose removal yields the steepest probability-mass drop. Organization. Section 2 reviews the literature. Section 3 formalizes matrices, operators, and fixed-point computation. Section 4 defines LMS K /BRE K . Section 5 presents properties and bounds. Section 6 provides examples and evidence. 2. Background and Related Work Account-based pivots vs. vulnerability-driven traversal. Beyond exploit-based movement (e.g., T1210) [ MITRE 2025d ] , ATT&CK catalogs techniques enabling lateral movement without software flaws: Valid Accounts (T1078) and Remote Services (T1021, including RDP/SSH) describe abusing legitimate credentials to authenticate and operate on remote hosts [ MITRE 2025f ; MITRE 2025b ; MITRE 2025c ; MITRE 2025e ] . Our formulation treats interactive services as high-pivot edges when paired with account access, independent of CVE posture. Attack graphs and probabilistic security metrics. Classical attack graphs model preconditions/exploits and enumerate paths to goals [ Phillips and Swiler 1998 ; Sheyner et al. 2002 ] . Later work augments graphs with probabilities and costs. We deliberately do not encode per-CVE logic; instead we treat reachable services as coarse pivot opportunities whose probabilities compose multiplicatively with noisy-OR unions, trading granularity for deployability on routine network snapshots. Algebraic path problems and semiring views. Many path computations admit uniform treatment via algebraic path frameworks replacing (+,\times) by problem-specific (\oplus,\otimes) . We adopt multiplication for per-step success and probabilistic sum (noisy-OR) for unions, yielding a monotone iteration toward a least fixed point; this aligns with semiring-style closures, though our \oplus is non-idempotent [ Mohri 2002 ; Sanmartín and Naranjo 2022 ] . Noisy-OR and causal independence. Our union operator is the standard noisy-OR from probabilistic graphical models, aggregating independent causal mechanisms with low elicitation burden [ Heckerman 1994 ] . Generalizations (e.g., Noisy-MAX) clarify when such factorizations are appropriate. Epidemic spreading analogies. Multiplicative transmission along edges with union across parallel paths is mathematically analogous to network contagion; classic results connect topology to spread [ Newman 2002 ] . Our aim is risk measurement (expected reach/blast), not epidemic prediction, but the analogy explains the outsized effect of a small set of high-pivot edges. AD attack-path tooling. In enterprise Windows domains, BloodHound operationalizes graph-based lateral-movement analysis at scale [ SpecterOps 2025 ] . We follow similar instincts (directed, typed edges) but compute a probability-weighted closure and provide system-level metrics for control evaluation. Zero Trust and ransomware guidance. NIST SP 800-207 sets the architectural principle; ransomware advisories emphasize limiting RDP/remote services to reduce both initial access and lateral movement [ Rose et al. 2020 ; CISA 2025a ; CISA 2025b ; CISA 2025c ] . 2.1 Graph-theoretic underpinnings k -hop walks via adjacency powers. Let A be the (directed) adjacency matrix of the network. Classical algebraic graph theory shows that (A^{k})_{ij} equals the number of i{\to}j walks of length k . 1 1 1 As usual, walks may repeat vertices; paths are walks with no repeated vertices. Our probabilistic closure works with walks. See, e.g., Godsil and Royle 2001 ; Biggs 1993 ; Diestel 2005 . This formalizes the “ k -hop” notion we use when bounding series A+AW+\cdots+AW^{K-1} . Boolean-OR reachability and transitive closure. Over the Boolean semiring (\lor,\land) , the transitive closure (reachability) can be computed by Boolean matrix powering and union, R\;=\;A\lor A^{(2)}\lor\cdots\lor A^{(n)}\,, or equivalently via Warshall’s dynamic program for transitive closure [ Warshall 1962 ] .