On controller performance in software-defined networks
Bibliographic record
Abstract
Network architectures in which the control plane is decoupled from the data plane have been growing in popularity. Among the main arguments for this approach is that it provides a more structured software environment for developing network-wide abstractions while potentially simplifying the data plane. As has been adopted elsewhere [11], we refer to this split architecture as Software-Defined Networking (SDN). While it has been argued that SDN is suitable for some deployment environments (such as homes [17, 13], data centers [1], and the enterprise [5]), delegating control to a remote system has raised a number of questions on control-plane scaling implications of such an approach. Two of the most often voiced concerns are: (a) how fast can the controller respond to data path requests?; and (b) how many data path requests can it handle per second? There are some references to the performance of SDN systems in the literature [16, 5, 3]. For example, an oft-cited study shows that a popular network controller (NOX) handles around 30k flow initiation events1 per second while maintaining a sub-10ms flow install time [14]. Unfortunately, recent measurements of some deployment environments suggests that these numbers are far from sufficient. For example, Kandula et al. [9] found that a 1500-server cluster has a median flow arrival rate of 100k flows per second. Also, Benson et al. [2] show that a network with 100 switches can have spikes of 10M flows arrivals per second in the worst case. In addition, the 10ms flow setup delay of an SDN controller would add a 10% delay to the majority of flows (short-lived) in such a network. This disconnect between relatively poor controller performance and high network demands has motivated a
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.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.006 | 0.034 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.001 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.003 | 0.006 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.004 | 0.001 |
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.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".