FIRST mermaid graph.md

FIRST mermaid graph

Keeping in mind the "Latency Numbers Every Programmer Should Know", i.e. network access is expensive.

Online request (service is provided by hashlookup online API)

• existing / not existing have the same high cost
• network access takes a long time
• leak every hash to hashlookup service

sequenceDiagram
    Alice->>Hashlookup: Do you have $HASH?
    alt is existing 
        Hashlookup->>Alice: details($HASH)
    else is not existing 
        Hashlookup->>Alice: $HASH not existing
    end

Offline request (service is provided by a bloom filter file)

• existing / not existing have the same low cost
• there is a probability of false positive
• no details available

sequenceDiagram
    Alice->>Filter: Do you have $HASH?
    alt seems existing 
        Filter->>Alice: $HASH is probably existing
    else is not existing 
        Filter->>Alice: $HASH not existing
    end

Hybrid request (service is provided by a bloom filter file and the online API)

• existing / not existing have a low cost
• getting hash's detail still have a high cost but occurs less often
• there is a probability of hitting some false positives where the filter says a %HASH is in the dataset but a query on the service yields no details.

sequenceDiagram
    Alice->>Filter: Do you have $HASH?
    alt seems existing 
        Filter->>Alice: $HASH is probably existing
        Alice->>Hashlookup: Do you have $HASH?
        alt is existing 
            Hashlookup->>Alice: details($HASH)
        else is not existing (False Positive)
            Hashlookup->>Alice: $HASH not existing
        end
    else is not existing 
        Filter->>Alice: $HASH not existing
    end