MétaCan
Menu
Back to cohort
Record W6944540845 · doi:10.21227/v32a-2c18

CRAWDAD icsi/netalyzr-android

2022· dataset· en· W6944540845 on OpenAlex

Why this work is in the frame

A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.

aboutThe title or abstract carries a Canadian signal from the geographic lexicon.
no affNo Canadian affiliation: this work is invisible to an affiliation-only frame.
No Canadian affiliation. An affiliation-only frame, the usual design, would never have seen this work. It is one of the works that make the case for inverting the frame.

Bibliographic record

VenueIEEE DataPort · 2022
Typedataset
Languageen
Field
Topic
Canadian institutionsnot available
Fundersnot available
KeywordsAndroid (operating system)TroubleshootingBackupMobile broadbandData qualityRoamingUser experience designData collectionPersonalization

Abstract

fetched live from OpenAlex

Mobile data collected using the Netalyzr for Android App.This dataset was collected by the ICSI Netalyzr app for Android to develop a characterization of how operational decisions, such as network configurations, business models, and relationships between operators introduce diversity in service quality and affect user security and privacy. We delve in detail beyond the radio link and into network configuration and business relationships in six countries. We identify the widespread use of transparent middleboxes such as HTTP and DNS proxies, analyzing how they actively modify user traffic, compromise user privacy, and potentially undermine user security. In addition, we identify network sharing agreements between operators, highlighting the implications of roaming and characterizing the properties of MVNOs, including that a majority are simply rebranded versions of major operators. More broadly, our findings using this data highlight the importance of considering higher-layer relationships when seeking to analyze mobile traffic in a sound fashion.date/time of measurement start: 2013-10-22date/time of measurement end: 2014-09-01collection environment: This dataset is collected using the Netalyzr for Android app. This app is available for free from the Google Play website for anyone to install and run. We analyzed data for a 9 month period from six countries: US, CA, UK, FR, DE, and AU.network configuration: Android phones connected through 3G and 4G networks. Rooted and unrooted devices, and multi-user/multi-device.data collection methodology: The data was collected by crowd-sourcing means. Users run proactively Netalyzr for Android App  to troubleshoot their network configuration or understand their network and how it behaves. No private data is collected without user's consent. Google Play link to the app: https://play.google.com/store/apps/details?id=edu.berkeley.icsi.netalyzr.androidsanitization: The dataset contains exclusively the sessions used for the core of the paper (MNO and MVNO characterization in the USA, Canada, France, Germany, Great Britain and Australia). We excluded users connected through VPNs, users with public IP addresses, users on international roaming, users connected through femtocells, users with customized network configurations (e.g. custom HTTP proxies and DNS resolvers), and sessions coming from engineering mode networks according to ITU. For the remaining Netalyzr sessions, we excluded sensitive fields such as passwords/usernames of APN settings, location information, base station information, and sensitive information injected on HTTP headers by proxies. IPv4 and IPv6 addresses are anonymized by performing a /16 and /32 sub-netting respectively. FQDNs are not also released as they contain information that can identify the users in many cases. For accessing a larger public Netalyzr dataset with more detailed values and all the collected variables, visit PREDICT: https://www.predict.orghole: Operator name, MCC/MNC values, as well as extra carrier information can be noisy or missing as a consequence of sessions generated by MVNO subscribers, network sharing agreements between operators, or even due to inconsistencies on Android's API (the dataset comprises handsets running versions from 2.2.3 to 5, which may also be modified by the vendor/mobile provider in their subsidized phones) or inaccurate APN settings on the handset (e.g. sometimes Android returns an empty MCC/MNC or an empty operator name). These sessions can be reconstructed. Other errors may appear on Netalyzr-specific tests (e.g. proxy detection and behavior characterization) due to connectivity problems or peculiar handset configurations. Our Mobisys'15 paper "Beyond the Radio: Illuminating the Higher Layers of Mobile Networks" contains further details about the data sanytization process, and the method followed for the study.error: It is a dataset collected through crowd-sourcing means. Caution is advised at the time of interpreting the data.limitation: Due to technical limitations, we cannot release an app for iOS, so this data is limited to Android users.note: Do not hesitate to contact us on netalyzr-help@icsi.berkeley.edu for questions.Traceseticsi/netalyzr-android/middleboxesDetails of middlebox behavior in cellular networks. The traceset contains a subset of the data collected from the Netalyzr for Android App.measurement purpose: Network Diagnosis, Network Performance AnalysisIP Addressing: Netalyzr identifies the client's local IP address via Android's APIs and system properties, and uses TCP connections and UDP flows to our echo servers to identify the public IP address of the device. We use the whois tool to identify the organization owning the IP address.Cellular Network Provider Identification: To identify the network service operator we use Android's TelephonyManager and ConnectivityManager APIs, and extract the APN settings as reported by the handset. This allows us to identify the name of the mobile operator, the name of the operator as reported by the SIM card, the APN providing the service, the cell ID (where users allow it), the 3GPP standard providing the service, as well as the MNC and MCC parameters.Location: Android allows us to extract city-level device location if the user allows it. This information is useful to identify where roaming happens between mobile operators, and identify locations with poor network performance.HTTP proxies.Non-responsive server test: TCP-terminating proxies may be deployed in cellular networks for performance improvement. Such proxies are likely to respond with a SYN-ACK to a client's connection request before connecting to the intended origin server. We test for this behavior by attempting a connection to a server that replies with a RST. If the Netalyzr client's attempt to connect to this server on port 80 initially succeeds, this indicates the presence of a TCP-terminating proxy.Header modification test: RFC 2616 specifies that systems should treat HTTP header names as case-insensitive, and, with few exceptions, free of ordering requirements. Furthermore, RFC 2615 indicates that any proxy must add the Via header to indicate its presence to intermediate protocols and recipients. Netalyzr fetches custom content from our server using mixed-cased request and response headers in a known order. Any changes indicate the presence a proxy. This method also allows identifying additional headers added by the HTTP proxy, as in the case of tracking headers, and whether intermediate proxies modify traffic using techniques such as image transcoding, which can affect the fidelity of content delivered to mobile clients through CDNs and other cloud infrastructure.HTTP enforcement test: In addition to standard HTTP, Netalyzr attempts to fetch an entity using the protocol declaration ICSI/1.1 instead of HTTP/1.1. If this request is rejected, we know that the network has a protocol-parsing proxy.Invalid Host header value test. CERT VU 435052 describes how some in-path proxies would interpret the Host request header and attempt to contact the listed host rather than forward the request to the intended address. We check for this vulnerability by fetching from our server with an alternate Host header of www.google.com. The presence of this vulnerability in commercial proxies is alarming as it suggests that operators may not have their middlebox software upgraded, potentially having other vulnerabilities not covered by our test suite.icsi/netalyzr-android/middleboxes Trace   middleboxes-trace: The data exposing middlebox (HTTP and DNS) behaviour in cellular networksconfiguration: Crowdsourced data collection using Netalyzr for Android appformat: The tuple (id,time,raw_op_name,clean_op_name,country,raw_cellular_technology,3gpp_family,mcc,mnc,apn,apn_name,extra_carrier_info,global_ip,ip_dns,ip_dns_proxy,ip_http_proxy,http_content_change,http_hdr_reorder,http_hdr_injection,invalid_host_name_vulnerability,http_enforcement,http_default_compression,transcoding,dns_direct_mangled,dns_direct_proxy,dns_direct_changed_id,roaming_indicator,rooted,http_header_injected_list  - id - integer- time - timestamp- raw_op_name - operator name as reported by Android's Telephony Manager- clean_op_name - operator name after applying our filter- country - device country as reported by android- raw_cellular_technology - 3GPP technology as reported by Android Connectivity/Telephony Manager- 3gpp_family - 3GPP family after applying our filter (i.e. UMTS/HSPA, LTE, CDMA)- mcc - Mobile Country Code. Asigned by ITU. Identifies the country. As reported by Android's Telephony Manager- mnc - Mobile Network Code. Asigned by ITU. Identifies the operator (generally the radio operator). As reported by Android Telephony Manager- apn - APN information (not all android devices return a value)- apn_name - APN Name (not all android devices return a value)- extra_carrier_info - Optionally supplied extra information about the- network state. Provided by Android Connectivity Manager- global_ip - Public IP address (/16 for IPv4 and /64 for IPv6)- ip_dns - IP address of the default DNS Resolver (as seen by Netalyzr)- ip_dns_proxy - Address of a DNS proxy ( as seen from Netalyzr server).- ip_http_proxy - IP address of the proxy in network ( as seen from Netalyzr server).- http_content_change - HTTP Content has been modified. Boolean not as reported by Android- rooted - Whether the phone is rooted or not (allows executing "su"). Security vulnerability.- http_header_injected_list - List of HTTP headers injected by the proxy.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.

Full frame distilled prediction

Teacher imitation

Not calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.

metaresearch head score (Codex)0.002
metaresearch head score (Gemma)0.000
Version: codex-gemma-dda1882f352aValidation status: machine_predicted_unvalidated
Candidate categoriesMeta-epidemiology (narrow), Research integrity, Insufficient payload (model declined to judge)
Consensus categoriesInsufficient payload (model declined to judge)
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Not applicable · Consensus signal: Not applicable
GenreCandidate signal: Dataset · Consensus signal: Dataset
Teacher disagreement score0.172
Threshold uncertainty score1.000

Codex and Gemma teacher scores by category

CategoryCodexGemma
Metaresearch0.0020.000
Meta-epidemiology (narrow)0.0010.001
Meta-epidemiology (broad)0.0010.000
Bibliometrics0.0010.001
Science and technology studies0.0000.000
Scholarly communication0.0000.001
Open science0.0040.001
Research integrity0.0010.002
Insufficient payload (model declined to judge)0.2410.069

Machine scores (provisional)

The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.

Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.

Opus teacher head0.031
GPT teacher head0.299
Teacher spread0.268 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it

Quick stats

Citations0
Published2022
Admission routes1
Has abstractyes

Explore more

Same venueIEEE DataPortFrench-language works237,207