hashlookup-lookup-strategies.md

hashlookup lookup strategies

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

The following diagrams show the different strategies to use for lookup while keeping latency or privacy in mind.

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

Loading

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

Loading

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 to exist 
        Filter->>Alice: $HASH probably exists
        Alice->>Hashlookup: Do you have $HASH?
        alt exists 
            Hashlookup->>Alice: details($HASH)
        else is not existing (False Positive)
            Hashlookup->>Alice: $HASH doesn't exist
        end
    else is not existing 
        Filter->>Alice: $HASH doesn't exist
    end

Loading