Reduce Serialized Size of HybridInfo (#1485)

This PR does a few things 1) Changes the codebase to assume helper_origin is a static string and does not need to be stored anywhere. It is removed from structs and outputs and simply grabbed from the place where it's defined if needed. 2) Additionally removes DOMAIN from serialized HybridInfo bytes. 3) Adds a second "to_bytes" method for HybridInfo structs. While we don't need hybrid_info and DOMAIN to deserialize HybridInfo, they are needed when we're actually encrypting/decrypting. For those purposes, a new to_enc_bytes() function is added.
private-attribution · Dec 8, 2024 · e7a234a · e7a234a
1 parent d30f67e
commit e7a234a
Show file tree

Hide file tree

Showing 10 changed files with 113 additions and 256 deletions.
diff --git a/ipa-core/src/cli/crypto/hybrid_decrypt.rs b/ipa-core/src/cli/crypto/hybrid_decrypt.rs
@@ -126,12 +126,8 @@ impl HybridDecryptArgs {
                     .reconstruct()
                     .as_u128();
                     let key_id = impression_report1.info.key_id;
-                    let helper_origin = impression_report1.info.helper_origin;
 
-                    writeln!(
-                        writer,
-                        "i,{match_key},{breakdown_key},{key_id},{helper_origin}"
-                    )?;
+                    writeln!(writer, "i,{match_key},{breakdown_key},{key_id}")?;
                 }
                 (
                     HybridReport::Conversion(conversion_report1),
@@ -154,12 +150,11 @@ impl HybridDecryptArgs {
                     .reconstruct()
                     .as_u128();
                     let key_id = conversion_report1.info.key_id;
-                    let helper_origin = conversion_report1.info.helper_origin;
                     let conversion_site_domain = conversion_report1.info.conversion_site_domain;
                     let timestamp = conversion_report1.info.timestamp;
                     let epsilon = conversion_report1.info.epsilon;
                     let sensitivity = conversion_report1.info.sensitivity;
-                    writeln!(writer, "c,{match_key},{value},{key_id},{helper_origin},{conversion_site_domain},{timestamp},{epsilon},{sensitivity}")?;
+                    writeln!(writer, "c,{match_key},{value},{key_id},{conversion_site_domain},{timestamp},{epsilon},{sensitivity}")?;
                 }
                 _ => {
                     panic!("Reports are not all the same type");

diff --git a/ipa-core/src/cli/crypto/hybrid_encrypt.rs b/ipa-core/src/cli/crypto/hybrid_encrypt.rs
@@ -300,14 +300,12 @@ mod tests {
 
     #[tokio::test]
     async fn try_encrypting_something() {
-        let helper_origin = "HELPER_ORIGIN".to_string();
         let conversion_site_domain = "meta.com".to_string();
         let records = vec![
             TestHybridRecord::TestConversion {
                 match_key: 12345,
                 value: 2,
                 key_id: 0,
-                helper_origin: helper_origin.clone(),
                 conversion_site_domain: conversion_site_domain.clone(),
                 timestamp: 100,
                 epsilon: 0.0,
@@ -317,7 +315,6 @@ mod tests {
                 match_key: 12345,
                 value: 5,
                 key_id: 0,
-                helper_origin: helper_origin.clone(),
                 conversion_site_domain: conversion_site_domain.clone(),
                 timestamp: 101,
                 epsilon: 0.0,
@@ -327,7 +324,6 @@ mod tests {
                 match_key: 23456,
                 breakdown_key: 4,
                 key_id: 0,
-                helper_origin: helper_origin.clone(),
             },
         ];
         let mut input_file = NamedTempFile::new().unwrap();

diff --git a/ipa-core/src/cli/csv.rs b/ipa-core/src/cli/csv.rs
@@ -29,24 +29,19 @@ impl Serializer for crate::test_fixture::hybrid::TestHybridRecord {
                 match_key,
                 breakdown_key,
                 key_id,
-                helper_origin,
             } => {
-                write!(
-                    buf,
-                    "i,{match_key},{breakdown_key},{key_id},{helper_origin}"
-                )?;
+                write!(buf, "i,{match_key},{breakdown_key},{key_id}")?;
             }
             crate::test_fixture::hybrid::TestHybridRecord::TestConversion {
                 match_key,
                 value,
                 key_id,
-                helper_origin,
                 conversion_site_domain,
                 timestamp,
                 epsilon,
                 sensitivity,
             } => {
-                write!(buf, "c,{match_key},{value},{key_id},{helper_origin},{conversion_site_domain},{timestamp},{epsilon},{sensitivity}")?;
+                write!(buf, "c,{match_key},{value},{key_id},{conversion_site_domain},{timestamp},{epsilon},{sensitivity}")?;
             }
         }
 

diff --git a/ipa-core/src/cli/playbook/input.rs b/ipa-core/src/cli/playbook/input.rs
@@ -62,9 +62,7 @@ impl InputItem for TestHybridRecord {
         let event_type = s.chars().nth(0).unwrap();
         match event_type {
             'i' => {
-                if let [_, match_key, number, key_id, helper_origin] =
-                    s.splitn(5, ',').collect::<Vec<_>>()[..]
-                {
+                if let [_, match_key, number, key_id] = s.splitn(4, ',').collect::<Vec<_>>()[..] {
                     let match_key: u64 = match_key
                         .parse()
                         .unwrap_or_else(|e| panic!("Expected a u64, got {match_key}: {e}"));
@@ -80,16 +78,15 @@ impl InputItem for TestHybridRecord {
                         match_key,
                         breakdown_key: number,
                         key_id,
-                        helper_origin: helper_origin.to_string(),
                     }
                 } else {
                     panic!("{s} is not a valid {}", type_name::<Self>())
                 }
             }
 
             'c' => {
-                if let [_, match_key, number, key_id, helper_origin, conversion_site_domain, timestamp, epsilon, sensitivity] =
-                    s.splitn(9, ',').collect::<Vec<_>>()[..]
+                if let [_, match_key, number, key_id, conversion_site_domain, timestamp, epsilon, sensitivity] =
+                    s.splitn(8, ',').collect::<Vec<_>>()[..]
                 {
                     let match_key: u64 = match_key
                         .parse()
@@ -118,7 +115,6 @@ impl InputItem for TestHybridRecord {
                         match_key,
                         value: number,
                         key_id,
-                        helper_origin: helper_origin.to_string(),
                         conversion_site_domain: conversion_site_domain.to_string(),
                         timestamp,
                         epsilon,

diff --git a/ipa-core/src/protocol/hybrid/agg.rs b/ipa-core/src/protocol/hybrid/agg.rs
@@ -196,20 +196,17 @@ pub mod test {
 
     #[allow(clippy::too_many_lines)]
     fn get_records() -> Vec<TestHybridRecord> {
-        let helper_origin = "HELPER_ORIGIN".to_string();
         let conversion_site_domain = "meta.com".to_string();
         let shard1_records = [
             TestHybridRecord::TestImpression {
                 match_key: SHARD1_MKS[0],
                 breakdown_key: 45,
                 key_id: 0,
-                helper_origin: helper_origin.clone(),
             },
             TestHybridRecord::TestConversion {
                 match_key: SHARD1_MKS[1],
                 value: 1,
                 key_id: 0,
-                helper_origin: helper_origin.clone(),
                 conversion_site_domain: conversion_site_domain.clone(),
                 timestamp: 102,
                 epsilon: 0.0,
@@ -219,7 +216,6 @@ pub mod test {
                 match_key: SHARD1_MKS[2],
                 value: 3,
                 key_id: 0,
-                helper_origin: helper_origin.clone(),
                 conversion_site_domain: conversion_site_domain.clone(),
                 timestamp: 103,
                 epsilon: 0.0,
@@ -229,7 +225,6 @@ pub mod test {
                 match_key: SHARD1_MKS[3],
                 value: 4,
                 key_id: 0,
-                helper_origin: helper_origin.clone(),
                 conversion_site_domain: conversion_site_domain.clone(),
                 timestamp: 104,
                 epsilon: 0.0,
@@ -239,19 +234,16 @@ pub mod test {
                 match_key: SHARD1_MKS[4],
                 breakdown_key: 1,
                 key_id: 0,
-                helper_origin: helper_origin.clone(),
             }, // duplicated impression with same match_key
             TestHybridRecord::TestImpression {
                 match_key: SHARD1_MKS[4],
                 breakdown_key: 2,
                 key_id: 0,
-                helper_origin: helper_origin.clone(),
             }, // duplicated impression with same match_key
             TestHybridRecord::TestConversion {
                 match_key: SHARD1_MKS[5],
                 value: 7,
                 key_id: 0,
-                helper_origin: helper_origin.clone(),
                 conversion_site_domain: conversion_site_domain.clone(),
                 timestamp: 105,
                 epsilon: 0.0,
@@ -263,13 +255,11 @@ pub mod test {
                 match_key: SHARD2_MKS[0],
                 breakdown_key: 56,
                 key_id: 0,
-                helper_origin: helper_origin.clone(),
             },
             TestHybridRecord::TestConversion {
                 match_key: SHARD2_MKS[1],
                 value: 2,
                 key_id: 0,
-                helper_origin: helper_origin.clone(),
                 conversion_site_domain: conversion_site_domain.clone(),
                 timestamp: 100,
                 epsilon: 0.0,
@@ -279,13 +269,11 @@ pub mod test {
                 match_key: SHARD2_MKS[2],
                 breakdown_key: 78,
                 key_id: 0,
-                helper_origin: helper_origin.clone(),
             }, // NOT attributed
             TestHybridRecord::TestConversion {
                 match_key: SHARD2_MKS[3],
                 value: 5,
                 key_id: 0,
-                helper_origin: helper_origin.clone(),
                 conversion_site_domain: conversion_site_domain.clone(),
                 timestamp: 101,
                 epsilon: 0.0,
@@ -295,13 +283,11 @@ pub mod test {
                 match_key: SHARD2_MKS[4],
                 breakdown_key: 90,
                 key_id: 0,
-                helper_origin: helper_origin.clone(),
             }, // attributed twice, removed
             TestHybridRecord::TestConversion {
                 match_key: SHARD2_MKS[5],
                 value: 6,
                 key_id: 0,
-                helper_origin: helper_origin.clone(),
                 conversion_site_domain: conversion_site_domain.clone(),
                 timestamp: 102,
                 epsilon: 0.0,
@@ -311,7 +297,6 @@ pub mod test {
                 match_key: SHARD2_MKS[6],
                 value: 7,
                 key_id: 0,
-                helper_origin: helper_origin.clone(),
                 conversion_site_domain: conversion_site_domain.clone(),
                 timestamp: 103,
                 epsilon: 0.0,

diff --git a/ipa-core/src/protocol/hybrid/oprf.rs b/ipa-core/src/protocol/hybrid/oprf.rs
@@ -231,19 +231,16 @@ mod test {
                     match_key: 12345,
                     breakdown_key: 2,
                     key_id: 0,
-                    helper_origin: "HELPER_ORIGIN".to_string(),
                 },
                 TestHybridRecord::TestImpression {
                     match_key: 68362,
                     breakdown_key: 1,
                     key_id: 0,
-                    helper_origin: "HELPER_ORIGIN".to_string(),
                 },
                 TestHybridRecord::TestConversion {
                     match_key: 12345,
                     value: 5,
                     key_id: 0,
-                    helper_origin: "HELPER_ORIGIN".to_string(),
                     conversion_site_domain: "meta.com".to_string(),
                     timestamp: 100,
                     epsilon: 0.0,
@@ -253,7 +250,6 @@ mod test {
                     match_key: 68362,
                     value: 2,
                     key_id: 0,
-                    helper_origin: "HELPER_ORIGIN".to_string(),
                     conversion_site_domain: "meta.com".to_string(),
                     timestamp: 102,
                     epsilon: 0.0,
@@ -263,13 +259,11 @@ mod test {
                     match_key: 68362,
                     breakdown_key: 1,
                     key_id: 0,
-                    helper_origin: "HELPER_ORIGIN".to_string(),
                 },
                 TestHybridRecord::TestConversion {
                     match_key: 68362,
                     value: 7,
                     key_id: 0,
-                    helper_origin: "HELPER_ORIGIN".to_string(),
                     conversion_site_domain: "meta.com".to_string(),
                     timestamp: 104,
                     epsilon: 0.0,