| name | infocom-topic-selection |
| description | Use when deciding whether a networking project belongs at IEEE INFOCOM or should be routed to SIGCOMM, NSDI, MobiCom, ICNP, or an IEEE/ACM networking journal (ToN, JSAC), distinguishing INFOCOM by its breadth, its analytical/optimization tradition alongside systems, and the modeling-vs-building axis. |
INFOCOM Topic Selection
Decide the venue before drafting. INFOCOM — the IEEE International Conference on Computer
Communications, the IEEE ComSoc networking flagship — is the broad, large-scale networking
venue. Its defining feature relative to its siblings is breadth plus a living analytical
tradition: a well-posed model with a theorem and a convincing evaluation is a first-class INFOCOM
paper, not a second-class one. A pure systems-building paper is welcome, but so is a
scheduling/optimization/queueing/game-theory result that SIGCOMM or NSDI would push to a journal.
The routing question that matters most
The decisive question is rarely "is this networking?" but "is the contribution a model/analysis, a
built system, or a measurement — and which community rewards that shape now?" INFOCOM is the safe
home for the analytical and the broad; SIGCOMM/NSDI pull toward built-and-measured systems with a
deployment story. Use the finer signals below and the live calendar to choose.
Sibling-venue routing table
| Signal in your project | Better home | Why |
|---|
| Modeling/optimization/algorithmic core (scheduling, resource allocation, queueing, game theory, network economics) with analysis + simulation | IEEE INFOCOM | Its analytical tradition treats theorems-with-evaluation as first-class |
| Broad networking contribution, ready now, with the next INFOCOM deadline nearer | IEEE INFOCOM | Large flagship; broad scope and reviewer pool |
| A built, deployed system with a measured Internet/datacenter deployment story | ACM SIGCOMM | Rewards architecture + real deployment; smaller, systems-first |
| A systems artifact whose case is a real implementation and hard performance numbers | USENIX NSDI | Networked-systems building, artifact-heavy |
| Wireless/mobile system with real devices, PHY/MAC + measurement | ACM MobiCom / MobiSys | Mobile-computing center of gravity |
| A focused protocol-design/analysis contribution | IEEE ICNP | Protocols venue; narrower than INFOCOM |
| Study too long or too incremental for a nine-page conference limit | IEEE/ACM ToN, IEEE JSAC | Journals with no page ceiling and a revise-and-resubmit cycle |
Contribution shapes INFOCOM rewards
- Analytical / optimization — a scheduling, resource-allocation, routing, caching, or
network-economics problem formulated cleanly, with structural results or an algorithm with
provable guarantees, validated in simulation. The queueing/optimization lineage is native here.
- Protocol / mechanism design — a new protocol or control mechanism (congestion control,
access, coordination) with analysis and evaluation on a simulator or modest testbed.
- Wireless and PHY/MAC — modeling and technique for wireless, cellular, 5G/6G, IoT, or
backscatter, often blending analysis with measurement.
- Measurement / empirical networking — a study of real network behavior with a dataset and
methodology a reviewer can scrutinize.
- Learning-for-networking — ML applied to a networking problem, where the networking lesson
(not the model) is the contribution (see the modeling-swap test).
The modeling-vs-building and model-swap tests
Two quick tests sharpen a borderline verdict:
- Modeling-vs-building axis: if the heart is a formulation and analysis (with simulation to
confirm), INFOCOM fits and SIGCOMM/NSDI may not. If the heart is a built system with a
deployment/measurement story, the systems venues fit better and an analysis-only INFOCOM framing
undersells it.
- Model-swap test: if the paper leans on a learner, ask whether the networking lesson survives
swapping the model. If not, the model is the contribution and an ML venue fits better; if yes, the
networking framing is real and INFOCOM welcomes it.
Evidence maturity, without the ladder cliché
Fit is necessary but not sufficient. A formulation with a theorem but no evaluation reads as
incomplete for the main track; a system evaluated only on a toy topology needs realistic scale
before it convinces; a measurement without a methodology a reviewer can audit will not survive the
crowded pool. INFOCOM's scale means a marginal paper competes against ~1,400 others for ~19% of
slots — submit when the analysis and the evaluation are both ready.
Cheap reconnaissance before committing
[Scope] scan the last two INFOCOM programs (dblp conf/infocom) for your subarea
-> several recent papers = a reviewer pool exists; near-zero = mismatch or opening
[Shape] is your core a model/algorithm, a built system, or a measurement?
-> model/algorithm leans INFOCOM; built-and-deployed leans SIGCOMM/NSDI
[Calendar] compare the next INFOCOM deadline (late July) with SIGCOMM/NSDI/MobiCom/journal dates
-> route to the nearest honest fit rather than idling a year
Decision procedure
[Audience] who acts differently if the claim holds? -> network designers/operators/theorists?
[Claim type] analytical / protocol / wireless / measurement / learning-for-networking
[Axis] model-and-analyze -> INFOCOM; build-and-deploy -> SIGCOMM/NSDI; wireless system -> MobiCom
[Sibling check] protocol-focused -> ICNP; too long/incremental -> ToN/JSAC
[Verdict] INFOCOM main track / sibling venue / journal, with a one-line reason
Run this before the writing skills; a wrong venue decision wastes every later step. When the verdict
is INFOCOM, continue with infocom-workflow for the calendar and infocom-writing-style for the
paper shape.
