Second Community Review of Dcent - NC Allocator

[cryptowhizzard]

Latest report:
https://compliance.allocator.tech/report/f03018494/1730680343/report.md

Since the allocator has been running, there have been a total of 5 client application requests.

Overall, the SP Spark retrieval success rate has been good. We are in transition period from Lotus-miner to curio and working to improve retrieval further.

Thanks!

Tasks

Beta Give feedback

No tasks being tracked yet.

[filecoin-watchdog]

@cryptowhizzard
Allocator Application  
Compliance Report 
1st Review 
1st Review score: 10PiB granted

9.75 PiB granted to clients:

Client Name	DC	status
ChildMind Institute	250TiB	New
NOAAWater-Column Sonar Data Archive	1PiB	Existing
HumanPangenome Reference Consortium	5.5PiB	Existing
Hubbletelescope full dataset	100TiB	New
DistributedArchives for Neurophysiology Data Integration (DANDI)	2.75PiB	New
CIDgravity	250TiB	Existing

CIDgravity

• The client declared cooperation with 200 SPs (so far, made deals with 20).
• The retrieval shows some SPs with 0% (5 SPs); however, the majority of the SPs have a bearable retrieval level, oscillating between 30-95%.
• The client declared 8 replicas. However, there are already 12. 
• The client explained CID sharing.

ChildMind Institute

• Did the allocator perform the KYC?
• There is no data sample to browse.
• No SP IDs were provided in the form.
• This client used only 60 out of 250TiB and made deals with only 2 SPs, both in the same region. 
• Not enough information on data preparation.

NOAAWater-Column Sonar Data Archive

• Same applicant name as in the previous example of ChildMind Institute.
• There is no data sample to browse.
• No SP IDs were provided in the form.
• Not enough information on data preparation.
• The allocator indicated in the comment that the KYC was also performed for the client and SPs. 
• The client has only 3 SPs, all in the same region. 
• CID sharing occurred.
• 1 SP sealed almost 60% of the data. 

Hubbletelescope full dataset 

• Same applicant name as in previous examples of ChildMind institute and NOAAWater-Column Sonar Data Archive.
• Only half of the DC was used.
• The application was closed due to several issues. 

HumanPangenome Reference Consortium 

• Same applicant name as in previous examples of ChildMind institute, NOAAWater-Column Sonar Data Archive and Hubbletelescope full dataset.
• Similar issues as before: no data sample, lack of information on data preparation.
• Poor retrieval rate.
• Not the best replication. The report says," 68.91% of deals are for data replicated across less than 4 storage providers."
• This dataset was stored before on the filecoin: https://github.com/search?q=repo%3Afilecoin-project%2Ffilecoin-plus-large-datasets+Human+Pangenome&type=issues
• According to the report (taking into account only the current round), 2 allocations were awarded directly, bypassing allocator.tech. 
• The allocation schedule for this application is unclear to me. I'd love to hear some clarification.

DistributedArchives for Neurophysiology Data Integration (DANDI) 

• Same applicant name as in previous examples of ChildMind institute, NOAAWater-Column Sonar Data Archive, Hubbletelescope full dataset and HumanPangenome Reference Consortium.
• Similar issues as before: no data sample, lack of information on data preparation.
• CID sharing occurred.
• According to the report (considering only the current round), one allocation was awarded directly, bypassing allocator.tech. 
• With only 4 SPs, 2 sealed over 30% of total datacap.
• This dataset was stored before on the filecoin:
  https://github.com/search?q=repo%3Afilecoin-project%2Ffilecoin-plus-large-datasets+DANDI&type=issues

---

Additional questions:

• How is KYC performed?
• Where does the main communication with clients take place?
• What are your guidelines for determining the allocation schedule?

A few notes:

• This allocator leaves little additional information in the applications for future reviews.
• I wonder about this allocator's allocation schedule because it is not entirely consistent with its application. However, considering request sizes, I would like to hear the allocator's opinion on this first.
• The allocator gives much credit to its primary client, based, as I understand it, on the fact that they know each other in person.