Pith. sign in

REVIEW 4 major objections 6 minor 53 references

Automating Conflict-Aware ACL Configurations with Natural Language Intents

T0 review · 4 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

Pith's one-line read This paper argues that network operators can move from plain-language intents to complete, conflict-free, deployment-optimized ACL rules without hand-crafting prefixes and ports.

desk verdict A credible LLM-to-ACL pipeline that solves a real NetOps pain point, but the headline claims outrun the synthetic evaluation's path model. read the letter →

arxiv 2508.17990 v1 pith:EB76MJZY submitted 2025-08-25 cs.NI cs.AI

classification cs.NIcs.AI
keywords ACLconfigurationnaturallanguageintentslargemodelsSemantics-NetworkMappingTabletruly-matchedflowsEquivalentIntentSetconflictdetectiondeploymentoptimization
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper argues that configuring access control lists (ACLs) can be automated end-to-end from a plain-language intent: a large language model turns the intent into concrete rules, a conflict checker filters out false alarms caused by rule order and impossible paths, and an optimizer picks the fewest rule insertions that still enforce every intent. If this is right, a task that can occupy network operations teams for days would shrink to a pipeline measured in minutes, catching conflicts before they cause outages instead of fixing them after users report them. The system, called Xumi, reports roughly 90–98.5% intent comprehension accuracy before any manual correction in campus and cloud networks, over a 10x end-to-end speedup, and about 40% fewer deployed rules.

What carries the argument

The load-bearing mechanisms are three named constructs. Truly-Matched Flows (TMF) gives each rule its actual jurisdiction: $T_k = R_k - \bigcup_{i<k} T_i$, so conflicts are judged only on flows the rule really decides, not flows shadowed by earlier entries. Interface-path validation then asks whether a flagged conflict flow has any feasible route through that interface. Equivalent Intent Sets (EIS) gather all intents that could be satisfied by the same rule insertion given existing same-action rules on an interface, and the deployment optimizer turns these into integer programs over interfaces and paths. The Semantics-Network Mapping Table supplies the network ground truth that prevents the LLM from hallucinating prefixes and gateways.

What would settle it

In a topology with ECMP load-balancing over eight equal-cost paths, feed Xumi only the four shortest paths and put an existing deny rule on an excluded path; then submit a permit intent whose traffic would traverse that path. If Xumi reports no conflict while live traffic is denied, the path-completeness assumption fails.

Watch

Extended reading notes

Core claim

Xumi's central claim is that ACL configuration has a complete automation path if three problems are solved jointly. First, comprehension: a large language model, grounded by a Semantics-Network Mapping Table that names endpoints, prefixes, and gateways, fills a structured ACL Intermediate Representation that network operators review. Second, conflict detection: for every existing rule at position k, the system computes the flows the rule truly governs by subtracting flows already matched by higher-priority rules, $T_k = R_k - \bigcup_{i<k} T_i$, and then validates that a candidate conflict's traffic can actually route through the interface; this removes the false positives that make naive overlap detection roughly 70% less accurate. Third, deployment optimization: grouping intents by equivalent coverage and choosing bottleneck interfaces for deny rules minimizes total rule additions. With these components, the paper reports complete configuration of 20 cloud-network intents in about five minutes, conflict detection on hundreds of ACLs, and roughly 38.8% fewer deployed rules than baseline practice.

Load-bearing premise

Xumi assumes it knows the routing paths every flow will take; if real routing (for example, ECMP or BGP changes) sends traffic over paths outside the assumed set, conflicts will be missed and the deployment plan may place rules on the wrong interfaces.

Editorial extensions

If this is right

  • An operator can submit an intent like 'block ChatGPT in exam areas during finals' and receive complete ACL rules plus a deployment plan, with the human role reduced to approving the LLM's output.
  • Real conflicts are distinguished from false positives caused by shadowed rules and infeasible paths, so problems can be fixed before deployment rather than after users complain.
  • Fewer rules are inserted, about 40% fewer than today's practice, because deny rules can be placed at bottleneck interfaces and one rule can cover multiple intents with existing rules.
  • On a 171-router cloud network, 20 intents are configured end-to-end in about five minutes, and even an extreme-scale network finishes conflict detection in at most two hours.
  • The same translate-conflict-optimize workflow is proposed for future routing configuration automation, not just ACLs.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The subtraction idea behind truly-matched flows should transfer to any policy engine with ordered rules, such as firewall chains, security groups, or routing policies, where the effective rule for a packet is the first match.
  • A direct stress test would replace the static shortest-path assumption with live routing state from ECMP or BGP and measure how many reported conflicts change; that gap would quantify how much of Xumi's accuracy depends on path completeness.
  • The paper's own trend of larger savings at higher conflict ratios and larger topologies suggests the roughly 40% rule-reduction figure is a floor for dense, highly conflicted networks rather than a ceiling.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. The paper presents Xumi, an LLM-based system that automates ACL configuration from natural language intents in three stages. Intent Comprehension uses a Semantics-Network Mapping Table (SNMT) and hallucination-mitigating prompts to translate intents into an intermediate representation and then into concrete ACL rules. Conflict Detection & Resolution computes truly-matched flows (TMF) to avoid false positives caused by preceding rules and performs interface-path validation to filter conflicts that do not lie on feasible routing paths; conflicts are then resolved via operator-supplied 'protect' intents. Deployment Optimization formulates set-cover-style integer programs that exploit bottleneck deployment and equivalent intent sets to minimize the total number of added rules while satisfying all intents. The system is evaluated on three synthetic networks (41, 171, and 1026 routers) with four LLMs, reporting comprehension accuracy up to 98.5% without feedback, conflict detection accuracy 3.33x above the blind-overlap baseline, ~40% rule-additions reduction, and an end-to-end speedup of over 10x versus the baseline pipeline.

Significance. If the reported results hold under realistic conditions, Xumi would be a substantial step toward intent-based network management, reducing the manual effort of ACL translation, conflict resolution, and deployment planning. The paper's genuine strengths are the TMF concept for accurate conflict detection, the equivalent-intent-set formulation for deployment optimization, the careful handling of flow-set operations with bitarray encodings, and the broad ablation study across network scales and LLMs. The prompts are open-sourced, which is commendable for reproducibility. However, the central correctness claims are conditioned on an untested routing model, the evaluation is entirely synthetic, and half of the comprehension test intents are generated by the same model family used as the best test model. These issues require additional work before the headline claims can be accepted as stated.

major comments (4)
  1. [§8.1, §9, Algorithm 1, §6.2 Constraints (4)–(5)] The correctness of both conflict detection and deployment optimization is tied to the path set returned by Routing(...) in Algorithm 1, line 3, which in every experiment is instantiated as k-shortest paths with k=4 (§8.1). Section 9 states that routing is assumed 'correct and predefined.' In real deployments, ECMP, BGP policy, and traffic engineering produce path sets that can differ from a static top-k set; if a conflicting flow uses a path outside the model, Algorithm 1 line 4 returns False and the conflict is silently dropped, and Constraints (4)–(5) can yield a deployment plan that leaves an actual path unprotected. Because the evaluation never varies k, never compares against routing-table-derived paths, and never measures the sensitivity to path-set errors, the abstract's 'all potential conflicts' and the ~40% rule-reduction claims are not established for realistic routing. Please either validate against realistic routing tables, perform a sensitivity analysis over the path model (including both fewer and more paths), or explicitly re-frame the correctness claims as holding 'under the assumed routing model.'
  2. [§8.1, §8.3, Table 3] The comprehension accuracy numbers are measured on a dataset of 100 intents per network, of which 50 are manually crafted and 50 are generated by GPT-4o (§8.1). Reported accuracy, such as the 90% CloudNet figure with GPT-4o, is a joint number over both groups; since the same model family generated half of the test set, the results may be inflated by self-generation bias. Please report accuracy separately for manually crafted and LLM-generated intents, and for protect intents, across all four LLMs, so that the reader can assess the size of this effect.
  3. [§8.4] Conflict detection accuracy is defined as 'the proportion of existing rules where the correct conflict flow set is identified for the new intent rule.' Because the ground-truth conflict sets are computed under the same k-shortest path model that Xumi uses internally, the reported 3.33x improvement over the blind-overlap baseline demonstrates internal consistency of the TMF and path-validation logic rather than robustness to realistic routing. The paper should additionally report recall-oriented metrics, such as the number of missed conflict flows or conflict intents, and explicitly analyze the causes of any false negatives, especially those that arise from paths not present in the assumed routing model.
  4. [§8.2, Table 3] The headline 'over 10x' acceleration in Table 3 depends on the Baseline comprehension time of about 237 seconds for 20 intents, which is reported as coming from 'a simple user study with NetOps' (§8.2). No protocol, participant count, task description, or variance is provided, and this single number is used to compute the primary speedup claim. Please either report the full study details or replace this estimate with a reproducible measurement, for example by running the baseline comprehension loop without the hallucination mitigators on the same intent dataset.
minor comments (6)
  1. [§8.4] The text contains a typo: 'Figrue' should be 'Figure.'
  2. [§5.1] The displayed formula for T_k contains a formatting artifact ('T1 =R11'); please typeset the initialization and the set-difference expression clearly.
  3. [§6.2] The logical symbols in Constraints (2), (4), and (5) are rendered as corrupted glyphs ('Ü', 'Ô', 'Û'); please replace them with standard OR and AND notation, and check that the logical structure is unambiguous.
  4. [§6.2] The phrase 'the rule count for each resolved intent RI_i donated as t_i' should read 'denoted as t_i.'
  5. [Figure 2] The labels in Figure 2 (parts b–d) are cramped and hard to read, especially the deployment-interface annotations; please enlarge the fonts and spacing for camera-readiness.
  6. [§8.3] The phrase 'inExtremeNet' is missing a space; it should read 'in ExtremeNet.'

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: Xumi's pipeline is a genuine synthesis; the reported gains come from solving stated detection and optimization problems against explicit baselines, not from fitted inputs renamed as predictions.

full rationale

Xumi's derivation chain is self-contained and not circular. The three modules are disjoint: Intent Comprehension is an LLM prompting pipeline whose output is reviewed and approved by NetOps, and its accuracy is measured against the structured IR tuples used to construct each intent rather than against the pipeline's own output; the ablation results show genuine incremental value of CoT, few-shot, and self-reflection. Conflict Detection computes Truly-Matched Flows by set differences over existing ACLs and validates interfaces against known routing paths (Algorithm 1); the accuracy comparison against BO, BO+TMF, and BO+Path applies the same path model and existing-rule sets to all methods, so the reported 3.33x gain is an algorithmic comparison, not a reduction of the measured quantity to the method's own assumptions. Deployment Optimization is an explicit integer program (Eqs. 1-5) whose objective is rule additions and whose constraints encode intent satisfaction via the equivalent-intent-set definition; the ~38.8% rule reduction is the optimization gap against deliberately simpler baselines (Endpoint, Catch-All, Bottleneck), not a value fitted to the reported outcome in advance. The only self-citation with overlapping authors is [19] (NAssim) in the Related Work section, used solely to support the peripheral assumption that vendor configuration templates exist; it is not load-bearing for any central claim, and no uniqueness theorem or ansatz is imported from the authors' prior work. The acknowledged limitation in Section 9 that routing is assumed 'correct and predefined,' together with the evaluation's use of k-shortest paths with k=4, is an external-validity and correctness risk (real routing paths outside the assumed set could cause missed conflicts or suboptimal placements), but it is not a circularity: the conflict-detection and deployment logic do not define their results in terms of the same path set they claim to predict. No fitted parameter is renamed as a prediction, and the central claims do not reduce by construction to any input of the system.

Assumptions & free parameters 1 free parameters · 5 assumptions · 0 invented entities

The system's correctness depends on standard ACL semantics (top-down first-match), the modeled routing path set, the mutual conflict-freeness of new intents, and the accuracy of the SNMT. The evaluation additionally assumes default-rule policies per network type. None of these are derived in the paper; they are inputs from domain practice. There are no fitted free parameters; the k=4 path setting is an experimental hyperparameter, not a constant fitted to the reported metrics. No new physical entities are postulated; TMF, EIS, and the 'protect' action are system-internal definitions and data structures with no independent falsifiable handles outside the paper.

free parameters (1)
  • k (number of shortest paths used to model feasible routing paths) = 4
    Set in §8.1 as 'k-shortest paths with k=4'; all conflict validation and deny-path deployment constraints use this path set, so the reported detection accuracy and rule-reduction numbers depend on this hand-chosen value.
assumptions (5)
  • standard math ACL rules are processed top-down and the first matched rule's action applies.
    Used in TMF definition and throughout §5.1; standard ACL semantics from vendor docs [3,7,8].
  • domain assumption Feasible routing paths between endpoints are known to Xumi and are modeled as k-shortest paths with k=4 in the evaluation.
    Algorithm 1's interface-path validation and the deny-deployment path constraints in §6.2 both rely on this path set; the paper acknowledges in §9 that routing is assumed correct and predefined.
  • domain assumption New intents within the same configuration round are mutually conflict-free.
    Stated in §2.2 as 'without loss of generality'; this lets conflict detection ignore new-new interactions and lets deployment optimization treat resolved intents as independent units.
  • domain assumption The SNMT is accurate and timely for all entity-to-prefix and gateway mappings.
    LLM comprehension (§4.1) and rule/gateway generation (§4.4) assume the mapping table is correct; the paper says scripts build it from production metadata, but SNMT error rates are not evaluated.
  • domain assumption Each interface's ACL ends with a default rule (deny all for CampusNet, permit all for CloudNet/ExtremeNet).
    Stated in §5.1 footnote and §8.1; the TMF calculation includes this default rule and it affects conflict detection outcomes.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Automating Conflict-Aware ACL Configurations with Natural Language Intents." pith.science (2026). https://pith.science/paper/EB76MJZY

@misc{pith2026250817990,
  author       = {Pith},
  title        = {Pith review of: Automating Conflict-Aware ACL Configurations with Natural Language Intents},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EB76MJZY}},
  note         = {Machine review of arXiv:2508.17990}
}
read the original abstract

ACL configuration is essential for managing network flow reachability, yet its complexity grows significantly with topologies and pre-existing rules. To carry out ACL configuration, the operator needs to (1) understand the new configuration policies or intents and translate them into concrete ACL rules, (2) check and resolve any conflicts between the new and existing rules, and (3) deploy them across the network. Existing systems rely heavily on manual efforts for these tasks, especially for the first two, which are tedious, error-prone, and impractical to scale. We propose Xumi to tackle this problem. Leveraging LLMs with domain knowledge of the target network, Xumi automatically and accurately translates the natural language intents into complete ACL rules to reduce operators' manual efforts. Xumi then detects all potential conflicts between new and existing rules and generates resolved intents for deployment with operators' guidance, and finally identifies the best deployment plan that minimizes the rule additions while satisfying all intents. Evaluation shows that Xumi accelerates the entire configuration pipeline by over 10x compared to current practices, addresses O(100) conflicting ACLs and reduces rule additions by ~40% in modern cloud network.

Figures

Figures reproduced from arXiv: 2508.17990 by the authors.

Figure 1
Figure 1. Conflict detection. Naive method detecting flow overlaps between intents and existing rules with opposite actions can cause false positives. Xumi finally generates valid ACL rules from the approved IR for subsequent tasks (§4.4). 2.2 Conflict Detection & Resolution After Intent Comprehension, each intent can be characterized by a group of rule(s) that may need to be added to different ACLs. A new rule, however, may … view at source ↗
Figure 2
Figure 2. Deployment optimization methods of intent rules. is not on any path from DC1 to DC3. Therefore, this detected conflict has no actual effect and is a FP. Main Ideas. Xumi tackles FPs in conflict detection rigor￾ously. First, to avoid FPs caused by preceding rules, we define truly-matched flows to precisely characterize traffic matched to a rule at its position after all preceding rules fail in ACL. Conflicts are then… view at source ↗
Figure 3
Figure 3. Xumi architecture. cover 𝐼 with complementary rules, and group them into an equivalent intent set for that intent 𝐼 and interface 𝐽 (§6.1). Recently JINJING [46] also considers reducing rule addi￾tions in ACL updates, highlighting its importance in produc￾tion networks. However, it focuses on minimizing new rules on a single interface after the rules to be deployed are pre￾determined. For example, if “any→a deny”, “… view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Workflow of Intent Comprehension with an intent example. 4 Intent Comprehension Our goal is to instruct LLMs to understand NetOps’ natural language intents and derive ACL rules in quintuple format with their endpoint gateways with high accuracy, minimiz￾ing human invol…
Figure 5
Figure 5. Figure 5: Comprehension accuracy across prompts and LLMs. (a) shows accuracy without NetOps’ feedback across LLMs, (b) shows cumulative accuracy with feedback rounds using GPT-4o, and (c) evaluates specific hallucination mitigator under GPT-4o by excluding it from system prompt …
Figure 6
Figure 6. Figure 6: Conflict detection accuracy of Xumi and methods ignoring FPs. results of our 100-intent dataset, and we report the results on average over 5 runs. Detection Accuracy. As shown in Figrue 6, Xumi achieves overall superior detection accuracy compared to other meth￾ods, wi…
Figure 7
Figure 7. Figure 7: Normalized rule additions for different deployment optimization methods, with values normalized to corresponding baselines. it greatly reduces rule additions compared to bottleneck de￾ployment. As conflict ratios increase, it becomes more effec￾tive in reducing rule ad…
Figure 8
Figure 8. Figure 8: The skeleton of the prompts used in Intent Comprehension. Note that “Step 1–5” corresponds to the Chain-of-Thought Reasoning prompts, “Examples” represent the Few-Shot Demonstration prompts, and “Verify x.x” refers to the Self-Reflection prompts. Prompt Skeletons. As d…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

53 extracted references · 48 canonical work pages

  1. [1]

    Azure Global DC Network.https://azure.microsoft.com/en-us/explore/ global-infrastructure

  2. [2]

    https://azure.microsoft.com/en-us/products/ ai-services/openai-service/

    Azure OpenAI Service. https://azure.microsoft.com/en-us/products/ ai-services/openai-service/

  3. [3]

    https://www.cisco.com/c/en/us/td/docs/ios-xml/ios/ iproute_ospf/configuration/xe-16/iro-xe-16-book/iro-cfg.html

    Cisco ACL. https://www.cisco.com/c/en/us/td/docs/ios-xml/ios/ iproute_ospf/configuration/xe-16/iro-xe-16-book/iro-cfg.html

  4. [4]

    https://www.cisco.com/c/en/us/products/ios-nx-os- software/border-gateway-protocol-bgp/index.html

    Cisco BGP. https://www.cisco.com/c/en/us/products/ios-nx-os- software/border-gateway-protocol-bgp/index.html

  5. [5]

    https://apstra.com/products/

    Graph-Based Live Queries in AOS. https://apstra.com/products/

  6. [6]

    http://www.gurobi.com

    Gurobi Optimizer. http://www.gurobi.com

  7. [7]

    https://www.h3c.com/en/d_201907/1213007_294551_0.htm

    H3C ACL. https://www.h3c.com/en/d_201907/1213007_294551_0.htm

  8. [8]

    https://support.huawei.com/enterprise/en/doc/ EDOC1000178177/d8968298/configuring-a-basic-acl

    Huawei ACL. https://support.huawei.com/enterprise/en/doc/ EDOC1000178177/d8968298/configuring-a-basic-acl

Show all 53 references
  1. [9]

    https://huggingface.co/

    Hugging Face. https://huggingface.co/

  2. [10]

    https://huggingface.co/nvidia/Llama-3.1-70B- Instruct-FP8

    Llama3.1-70B-FP8. https://huggingface.co/nvidia/Llama-3.1-70B- Instruct-FP8

  3. [11]

    http://openconfig.net/

    OpenConfig. http://openconfig.net/

  4. [12]

    https://pypi.org/project/bitarray/

    Python Bitarray. https://pypi.org/project/bitarray/

  5. [13]

    https://huggingface.co/Qwen/Qwen2.5-72B- Instruct-GPTQ-Int4

    Qwen2.5-72B-Int4. https://huggingface.co/Qwen/Qwen2.5-72B- Instruct-GPTQ-Int4

  6. [14]

    https://anonymous.4open.science/r/Comprehension- opensource-17C8

    Xumi. https://anonymous.4open.science/r/Comprehension- opensource-17C8

  7. [15]

    https://developer.cisco.com/docs/nso/guides/the-yang- data-modeling-language/

    YANG Model. https://developer.cisco.com/docs/nso/guides/the-yang- data-modeling-language/

  8. [16]

    NetComplete: Practical Network-Wide Configuration Synthesis with Autocompletion

    Ahmed El-Hassany and Petar Tsankov and Laurent Vanbever and Martin Vechev. NetComplete: Practical Network-Wide Configuration Synthesis with Autocompletion. In Proc. USENIX NSDI, 2018

  9. [17]

    A Scalable, Commodity Data Center Network Architecture

    Al-Fares, Mohammad and Loukissas, Alexander and Vahdat, Amin. A Scalable, Commodity Data Center Network Architecture. InProc. ACM SIGCOMM, 2008

  10. [18]

    A General Approach to Network Configuration Verification

    Ryan Beckett, Aarti Gupta, Ratul Mahajan, and David Walker. A General Approach to Network Configuration Verification. In Proc. ACM SIGCOMM, 2017

  11. [19]

    Software-Defined Network Assimilation: Bridging the Last Mile Towards Centralized Network Configuration Management with NAssim

    Huangxun Chen, Yukai Miao, Li Chen, Haifeng Sun, Hong Xu, Li- bin Liu, Gong Zhang, and Wei Wang. Software-Defined Network Assimilation: Bridging the Last Mile Towards Centralized Network Configuration Management with NAssim. In Proc. ACM SIGCOMM, 2022

  12. [20]

    Detecting Vague Words & Phrases in Require- ments Documents in A Multilingual Environment

    Cruz, Breno Dantas and Jayaraman, Bargav and Dwarakanath, Anurag and McMillan, Collin. Detecting Vague Words & Phrases in Require- ments Documents in A Multilingual Environment. In 2017 IEEE 25th International Requirements Engineering Conference (RE) , 2017

  13. [21]

    Named Entity Recognition and Classi- fication in Historical Documents: A Survey

    Maud Ehrmann, Ahmed Hamdi, Elvys Linhares Pontes, Matteo Ro- manello, and Antoine Doucet. Named Entity Recognition and Classi- fication in Historical Documents: A Survey. 56(2):1–47, Sep. 2023

  14. [22]

    A General Approach to Network Configuration Analysis

    Fogel, Ari and Fung, Stanley and Pedrosa, Luis and Walraed-Sullivan, Meg and Govindan, Ramesh and Mahajan, Ratul and Millstein, Todd. A General Approach to Network Configuration Analysis. In Proc. USENIX NSDI, 2015

  15. [23]

    Fast Control Plane Analysis Using an Abstract Repre- sentation

    Aaron Gember-Jacobson, Raajay Viswanathan, Aditya Akella, and Ratul Mahajan. Fast Control Plane Analysis Using an Abstract Repre- sentation. In Proc. ACM SIGCOMM, 2016

  16. [24]

    Abandah and Ashraf Suyyagh and Mohammed Z

    Gheith A. Abandah and Ashraf Suyyagh and Mohammed Z. Khed- her. Correcting Arabic Soft Spelling Mistakes using BiLSTM-based Machine Learning. arXiv preprint arXiv:2108.01141, 2021

  17. [25]

    An Efficient System for Grammatical Error Correction on Mobile Devices

    Gothe, Sourabh Vasant and Dogra, Sushant and Chandra, Mritunjai and Sanchi, Chandramouli and Raja, Barath Raj Kandur. An Efficient System for Grammatical Error Correction on Mobile Devices. In 2021 IEEE 15th International Conference on Semantic Computing (ICSC) , 2021

  18. [26]

    Achieving High Utilization with Software-Driven WAN

    Chi-Yao Hong, Srikanth Kandula, Ratul Mahajan, Ming Zhang, Vi- jay Gill, Mohan Nanduri, and Roger Wattenhofer. Achieving High Utilization with Software-Driven WAN. InProc. ACM SIGCOMM, 2013

  19. [27]

    Jacobs, Ricardo J

    Arthur S. Jacobs, Ricardo J. Pfitscher, Rafael H. Ribeiro, Ronaldo A. Ferreira, Lisandro Z. Granville, Walter Willinger, and Sanjay G. Rao. Hey, Lumi! Using Natural Language for Intent-Based Network Man- agement. In Proc. USENIX ATC, 2021

  20. [28]

    B4: Experience with a Globally-Deployed Software Defined WAN

    Sushant Jain, Alok Kumar, Subhasree Mandal, Joon Ong, Leon Poutievski, Arjun Singh, Subbaiah Venkata, Jim Wanderer, Junlan Zhou, Min Zhu, Jon Zolla, Urs Hölzle, Stephen Stuart, and Amin Vah- dat. B4: Experience with a Globally-Deployed Software Defined WAN. In Proc. ACM SIGCOMM, 2013

  21. [29]

    Survey of Hallucination in Natural Language Generation

    Ji, Ziwei and Lee, Nayeon and Frieske, Rita and Yu, Tiezheng and Su, Dan and Xu, Yan and Ishii, Etsuko and Bang, Ye Jin and Madotto, An- drea and Fung, Pascale. Survey of Hallucination in Natural Language Generation. ACM Computing Surveys, 55(12):1–38, March 2023

  22. [30]

    Mistral 7B.arXiv preprint arXiv:2310.06825, 2023

    Jiang, Albert Q and Sablayrolles, Alexandre and Mensch, Arthur and Bamford, Chris and Chaplot, Devendra Singh and Casas, Diego de las and Bressand, Florian and Lengyel, Gianna and Lample, Guillaume and Saulnier, Lucile and others. Mistral 7B.arXiv preprint arXiv:2310.06825, 2023

  23. [31]

    CEGS: Con- figuration Example Generalizing Synthesizer

    Jianmin Liu and Li Chen and Dan Li and Yukai Miao. CEGS: Con- figuration Example Generalizing Synthesizer. In Proc. USENIX NSDI, 2025

  24. [32]

    The Nature of Data Center Traffic: Measurements & Analysis

    Kandula, Srikanth and Sengupta, Sudipta and Greenberg, Albert and Patel, Parveen and Chaiken, Ronnie. The Nature of Data Center Traffic: Measurements & Analysis. In Proc. ACM IMC, 2009

  25. [33]

    Semi-Oblivious Traffic Engineering: The Road Not Taken

    Praveen Kumar, Yang Yuan, Chris Yu, Nate Foster, Robert Kleinberg, Petr Lapukhov, Chiun Lin Lim, and Robert Soulé. Semi-Oblivious Traffic Engineering: The Road Not Taken. In Proc. USENIX NSDI, 2018

  26. [34]

    A Survey on Deep Learning for Named Entity Recognition

    Li, Jing and Sun, Aixin and Han, Jianglei and Li, Chenliang. A Survey on Deep Learning for Named Entity Recognition. IEEE Transactions on Knowledge and Data Engineering , 34(1):50–70, January 2020

  27. [35]

    Halueval: A Large-Scale Hallucination Eval- uation Benchmark for Large Language Models

    Li, Junyi and Cheng, Xiaoxue and Zhao, Wayne Xin and Nie, Jian- Yun and Wen, Ji-Rong. Halueval: A Large-Scale Hallucination Eval- uation Benchmark for Large Language Models. arXiv Preprint arXiv:2305.11747, 2023

  28. [36]

    Bond: Bert-Assisted Open-Domain Named Entity Recognition with Distant Supervision

    Liang, Chen and Yu, Yue and Jiang, Haoming and Er, Siawpeng and Wang, Ruijia and Zhao, Tuo and Zhang, Chao. Bond: Bert-Assisted Open-Domain Named Entity Recognition with Distant Supervision. In Proc. ACM SIGKDD, 2020

  29. [37]

    Automatic Detection of Vague Words and Sentences in Privacy Policies

    Logan Lebanoff and Fei Liu. Automatic Detection of Vague Words and Sentences in Privacy Policies. arXiv preprint arXiv:1808.06219, 2018

  30. [38]

    What do LLMs Need to Synthesize Correct Router Configurations? In Proc

    Mondal, Rajdeep and Tang, Alan and Beckett, Ryan and Millstein, Todd and Varghese, George. What do LLMs Need to Synthesize Correct Router Configurations? In Proc. ACM HotNets, 2023

  31. [39]

    Practical Intent-driven Routing Configuration Synthesis

    Sivaramakrishnan Ramanathan, Ying Zhang, Mohab Gawish, Yogesh Mundada, Zhaodong Wang, Sangki Yun, Eric Lippert, Walid Taha, Minlan Yu, and Jelena Mirkovic. Practical Intent-driven Routing Configuration Synthesis. In Proc. USENIX NSDI, 2023

  32. [40]

    Prompt Programming for Large Language Models: Beyond the Few-Shot Paradigm

    Reynolds, Laria and McDonell, Kyle. Prompt Programming for Large Language Models: Beyond the Few-Shot Paradigm. In Extended Ab- stracts of the 2021 CHI Conference on Human Factors in Computing Systems, 2021

  33. [41]

    Cantor Meets Scott: Semantic Foundations for Prob- abilistic Networks

    Steffen Smolka, Praveen Kumar, Nate Foster, Dexter Kozen, and Alexandra Silva. Cantor Meets Scott: Semantic Foundations for Prob- abilistic Networks. In Proceedings of ACM SIGPLAN Symposium on Principles of Programming Languages , 2017

  34. [42]

    An Information-Theoretic Approach to Prompt Engi- neering without Ground Truth Labels

    Sorensen, Taylor and Robinson, Joshua and Rytting, Christopher Michael and Shaw, Alexander Glenn and Rogers, Kyle Jeffrey and Delorey, Alexia Pauline and Khalil, Mahmoud and Fulda, Nancy and Wingate, David. An Information-Theoretic Approach to Prompt Engi- neering without Grou...

  35. [43]

    Probabilistic Verification of Network Configurations

    Samuel Steffen, Timon Gehr, Petar Tsankov, Laurent Vanbever, and Martin Vechev. Probabilistic Verification of Network Configurations. In Proc. ACM SIGCOMM, 2020. 14

  36. [44]

    Detecting Network Load Violations for Distributed Control Planes

    Kausik Subramanian, Anubhavnidhi Abhashkumar, Loris D’Antoni, and Aditya Akella. Detecting Network Load Violations for Distributed Control Planes. In Proceedings of ACM SIGPLAN Conference on Pro- gramming Language Design and Implementation , 2020

  37. [45]

    Robotron: Top-down Network Management at Facebook Scale

    Yu-Wei Eric Sung, Xiaozheng Tie, Starsky HY Wong, and Hongyi Zeng. Robotron: Top-down Network Management at Facebook Scale. In Proc. ACM SIGCOMM, 2016

  38. [46]

    Bingchuan Tian, Xinyi Zhang, Ennan Zhai, Hongqiang Harry Liu, Qiaobo Ye, Chunsheng Wang, Xin Wu, Zhiming Ji, Yihong Sang, Ming Zhang, Da Yu, Chen Tian, Haitao Zheng, and Ben Y. Zhao. Safely and Automatically Updating In-Network ACL Configurations with Intent Language. In Proc....

  39. [47]

    Llama: Open and Efficient Foundation Language Models

    Hugo Touvron, Thibaut Lavril, Gautier Izacard, Xavier Martinet, Marie- Anne Lachaux, Timothée Lacroix, Baptiste Rozière, Naman Goyal, Eric Hambro, Faisal Azhar, et al. Llama: Open and Efficient Foundation Language Models. arXiv preprint arXiv:2302.13971, 2023

  40. [48]

    Gomez, Lukasz Kaiser, and Illia Polosukhin

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Lukasz Kaiser, and Illia Polosukhin. Attention is All you Need. In Advances in Neural Information Processing Systems , 2017

  41. [49]

    NetConfEval: Can LLMs Facili- tate Network Configuration? Proceedings of the ACM on Networking , 2(CoNEXT2):1–25, June 2024

    Changjie Wang, Mariano Scazzariello, Alireza Farshin, Simone Ferlin, Dejan Kostić, and Marco Chiesa. NetConfEval: Can LLMs Facili- tate Network Configuration? Proceedings of the ACM on Networking , 2(CoNEXT2):1–25, June 2024

  42. [50]

    Making Network Configuration Human Friendly

    Wang, Changjie and Scazzariello, Mariano and Farshin, Alireza and Kostic, Dejan and Chiesa, Marco. Making Network Configuration Human Friendly. arXiv Preprint arXiv:2309.06342, 2023

  43. [51]

    Chain-of-Thought Prompting Elicits Reasoning in Large Language Models

    Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Fei Xia, Ed Chi, Quoc V Le, Denny Zhou, et al. Chain-of-Thought Prompting Elicits Reasoning in Large Language Models. In Advances in Neural Information Processing Systems, 2022

  44. [52]

    Step 1–5

    Duo Wu, Xianda Wang, Yaqi Qiao, Zhi Wang, Junchen Jiang, Shuguang Cui, and Fangxin Wang. NetLLM: Adapting Large Language Models for Networking. In Proc. ACM SIGCOMM, 2024. 15 Appendices A Comprehension Prompts and Human Interaction Example Figure 8. The skeleton of the prompts...

  45. [53]

    10.0.0.0/32→ 11.0.0.1/32 HTTP

    and “10.0.0.0/32→ 11.0.0.1/32 HTTP” (index 3), so its index set is {1, 3}, represented as the Bitarray “1010”. Using these precomputed index sets, we can efficiently compute the flow set for a rule. First, we enumerate all specifics of an attribute and take the union of their ...

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.