FOKSTRAUT and Samba--Dealing with Authentication and Performance Issues on a Large-scale Samba Service
Bibliographic record
Abstract
At the University of Alberta, we have approximately 55,000 user id's using central services authenticated by Kerberos. We use AFS for central file service. We use Samba to provide Windows compatible access to much of our central file service. Samba contains a number of useful features for Microsoft Windows compatibility, including a kludge to deal with the problem of Windows sending an all uppercase version of a user's password. We observed that when Windows connects to a share, it frequently attempts many incorrect passwords repeatedly before trying the correct one. This created a very heavy authentication load on our central Samba service when users would connect every morning and authenticate. We observed this load and noticed that most of our problems were caused by repeated attempts to authenticate, and the high cost of checking these attempts.To help reduce the load due to authentication, we implemented FOKSTRAUT, a set of modifications to Samba to cache recent password failures and successes in a DBM database built by the Samba server as it runs. By caching the recent failures we avoid expensive re-checks of the (many) other passwords Windows likes to send us. We also cache the correct case of the real password, and by doing so we avoid the expensive overhead of cracking an all uppercase password When Windows decides to send one. We also use FOKSTRAUT to cache the NT and LanMan password hashes of a users password once we see a successful authentication. This then allows us to use the newer Windows NT password hash after the user has connected once, without having to centrally convert and maintain a large SMB password file, and while maintaining the ability of our server to access services such as AFS which can not be authenticated against using the Windows password hash alone. Performance on our service has been drastically improved since the implementation of FOKSTRAUT.
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 | 0.024 |
| Meta-epidemiology (narrow) | 0.002 | 0.002 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.003 |
| Science and technology studies | 0.003 | 0.003 |
| Scholarly communication | 0.005 | 0.012 |
| Open science | 0.006 | 0.005 |
| Research integrity | 0.002 | 0.003 |
| Insufficient payload (model declined to judge) | 0.007 | 0.003 |
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".