MobiSys Related Work
Use this to audit novelty and coverage. Reopen the current CFP for dual-submission,
anonymity, and prior-publication rules before advising authors. At MobiSys the failure mode is
usually a too-narrow related-work section that misses a sibling-venue predecessor.
Positioning checks
- Separate the system contribution from an engineering improvement: a new runtime
mechanism, scheduling policy, on-device inference system, sensing service, or platform.
- Cover the mobile-systems lanes below; a bibliography that cites only machine-learning papers
tells a systems reviewer that known mobile-systems work may be getting rediscovered.
- Treat ACM DL, USENIX, and IEEE proceedings as archival unless current rules say otherwise.
- Cite arXiv and workshop versions in a way that preserves double-blind review; do not point
reviewers to identity-revealing pages.
- Use related work to sharpen what is new: a tighter energy budget, a lower latency tail, an
on-device capability that previously required the cloud, or a deployment others only
simulated.
Literature lanes to sweep
| Lane |
Typical venues |
What MobiSys reviewers check |
| Computation offload / edge |
MobiSys, MobiCom, NSDI, EuroSys |
Whether the nearest offload system is compared or distinguished |
| On-device ML systems |
MobiSys, SenSys, MLSys, ASPLOS |
Whether prior on-device runtimes/schedulers are acknowledged |
| Mobile OS / runtime |
MobiSys, OSDI, SOSP, EuroSys |
Whether the platform mechanism has an OS-systems predecessor |
| Mobile sensing / ubicomp |
MobiSys, SenSys, IMWUT |
Whether the sensing-service line is covered without misfiling |
| Energy / measurement |
MobiSys, SenSys, IMC |
Whether energy methodology follows established practice |
A related-work section that ignores the SenSys or OSDI predecessor of a mobile-systems idea is
a recognizable MobiSys reject pattern that no amount of on-device polish repairs.
Venue-verification discipline
The mobile-systems canon is easy to misattribute because MobiSys, MobiCom, and SenSys share
the SIGMOBILE umbrella:
- Verify each cited "MobiSys paper" on the ACM DL and dblp by matching the
conf/mobisys record; TaintDroid is OSDI, CenceMe is SenSys, RF-sensing classics are
MobiCom/SIGCOMM (see ../../resources/exemplars/library.md).
- Do not cite a paper as MobiSys from memory; a misattributed venue signals careless
scholarship to a specialist reviewer.
Concurrent-work judgment calls
- Independently concurrent arXiv work: cite neutrally, state the technical difference, and
avoid priority claims reviewers cannot verify.
- Your own workshop (e.g., HotMobile) version: typically non-archival and citable, but verify
against the current CFP and phrase the citation so double-blind review survives.
- When in doubt about the archival status of a venue, declare the overlap in the submission
form rather than gambling on a chair's interpretation.
Positioning vignette
Imagine the paper proposes a thermal-aware on-device inference runtime. Its nearest neighbors:
an on-device DNN scheduler at MobiSys with no thermal model, a mobile-GPU inference framework
at MobiSys tuned for peak not sustained load, and an OS-level DVFS mechanism at OSDI. The
novelty sentence should name all three contrasts — thermal-awareness where the scheduler had
none, sustained-load stability where the GPU framework optimized peak, and application-level
control where the OS mechanism was generic.
Output format
[Eligibility] clear / needs declaration / risky
[Closest lanes] <offload / on-device ML / mobile OS / sensing / energy>
[Nearest 3 works] <work -> distinction>
[Venue-verification risk] <none / misattribution issues>
[Novelty sentence] <MobiSys-ready contribution contrast>
1---2name: mobisys-related-work3description: Use when positioning a MobiSys submission against the mobile-systems literature — offload, on-device ML, mobile OS and runtimes, sensing services, and energy — covering the right lanes, handling concurrent work, verifying that cited "MobiSys papers" are MobiSys rather than MobiCom/SenSys/OSDI, and self-citing without breaking double-blind.4---56# MobiSys Related Work78Use this to audit novelty and coverage. Reopen the current CFP for dual-submission,9anonymity, and prior-publication rules before advising authors. At MobiSys the failure mode is10usually a **too-narrow** related-work section that misses a sibling-venue predecessor.1112## Positioning checks1314- Separate the **system contribution** from an engineering improvement: a new runtime15 mechanism, scheduling policy, on-device inference system, sensing service, or platform.16- Cover the mobile-systems lanes below; a bibliography that cites only machine-learning papers17 tells a systems reviewer that known mobile-systems work may be getting rediscovered.18- Treat ACM DL, USENIX, and IEEE proceedings as archival unless current rules say otherwise.19- Cite arXiv and workshop versions in a way that preserves double-blind review; do not point20 reviewers to identity-revealing pages.21- Use related work to sharpen what is new: a tighter energy budget, a lower latency tail, an22 on-device capability that previously required the cloud, or a deployment others only23 simulated.2425## Literature lanes to sweep2627| Lane | Typical venues | What MobiSys reviewers check |28|---|---|---|29| Computation offload / edge | MobiSys, MobiCom, NSDI, EuroSys | Whether the nearest offload system is compared or distinguished |30| On-device ML systems | MobiSys, SenSys, MLSys, ASPLOS | Whether prior on-device runtimes/schedulers are acknowledged |31| Mobile OS / runtime | MobiSys, OSDI, SOSP, EuroSys | Whether the platform mechanism has an OS-systems predecessor |32| Mobile sensing / ubicomp | MobiSys, SenSys, IMWUT | Whether the sensing-service line is covered without misfiling |33| Energy / measurement | MobiSys, SenSys, IMC | Whether energy methodology follows established practice |3435A related-work section that ignores the SenSys or OSDI predecessor of a mobile-systems idea is36a recognizable MobiSys reject pattern that no amount of on-device polish repairs.3738## Venue-verification discipline3940The mobile-systems canon is easy to misattribute because MobiSys, MobiCom, and SenSys share41the SIGMOBILE umbrella:4243- Verify each cited "MobiSys paper" on the **ACM DL** and **dblp** by matching the44 `conf/mobisys` record; TaintDroid is OSDI, CenceMe is SenSys, RF-sensing classics are45 MobiCom/SIGCOMM (see [`../../resources/exemplars/library.md`](../../resources/exemplars/library.md)).46- Do not cite a paper as MobiSys from memory; a misattributed venue signals careless47 scholarship to a specialist reviewer.4849## Concurrent-work judgment calls5051- Independently concurrent arXiv work: cite neutrally, state the technical difference, and52 avoid priority claims reviewers cannot verify.53- Your own workshop (e.g., HotMobile) version: typically non-archival and citable, but verify54 against the current CFP and phrase the citation so double-blind review survives.55- When in doubt about the archival status of a venue, declare the overlap in the submission56 form rather than gambling on a chair's interpretation.5758## Positioning vignette5960Imagine the paper proposes a thermal-aware on-device inference runtime. Its nearest neighbors:61an on-device DNN scheduler at MobiSys with no thermal model, a mobile-GPU inference framework62at MobiSys tuned for peak not sustained load, and an OS-level DVFS mechanism at OSDI. The63novelty sentence should name all three contrasts — thermal-awareness where the scheduler had64none, sustained-load stability where the GPU framework optimized peak, and application-level65control where the OS mechanism was generic.6667## Output format6869```text70[Eligibility] clear / needs declaration / risky71[Closest lanes] <offload / on-device ML / mobile OS / sensing / energy>72[Nearest 3 works] <work -> distinction>73[Venue-verification risk] <none / misattribution issues>74[Novelty sentence] <MobiSys-ready contribution contrast>75```
brycewang-stanford