Skip to content

Commit 47a4f36

Browse files
roloffsroloffs
committed
remote_exec.proto: add blob split and splice API
Depending on the software project, possibly large artifacts need to be downloaded from or uploaded to the remote CAS such as executables with debug information, comprehensive libraries, or even whole file system images. Such artifacts generate a lot of traffic when downloaded or uploaded. The blob-split API allows to split such artifacts into chunks at the remote side, to fetch only those parts that are locally missing, and finally to locally assemble the requested blob from its chunks. The blob-splice API allows to split such artifacts into chunks locally, to upload only those parts that are remotely missing, and finally to remotely splice the requested blob from its chunks. Since only the binary differences from the last download/upload are fetched/uploaded, the blob split and splice API can save a lot of network traffic between server and client.
1 parent 6c32c3b commit 47a4f36

File tree

1 file changed

+175
-0
lines changed

1 file changed

+175
-0
lines changed

‎build/bazel/remote/execution/v2/remote_execution.proto

+175
Original file line numberDiff line numberDiff line change
@@ -430,6 +430,93 @@ service ContentAddressableStorage {
430430
rpc GetTree(GetTreeRequest) returns (stream GetTreeResponse) {
431431
option (google.api.http) = { get: "/v2/{instance_name=**}/blobs/{root_digest.hash}/{root_digest.size_bytes}:getTree" };
432432
}
433+
434+
// Split a blob into chunks.
435+
//
436+
// This splitting API aims to reduce download traffic between client and
437+
// server, e.g., if a client needs to fetch a large blob that just has been
438+
// modified slightly since the last built. In this case, there is no need to
439+
// fetch the entire blob data, but just the binary differences between the two
440+
// blob versions, which are typically determined by deduplication techniques
441+
// such as content-defined chunking.
442+
//
443+
// Clients can use this API before downloading a blob to determine which parts
444+
// of the blob are already present locally and do not need to be downloaded
445+
// again. The server splits the blob into chunks according to a specified
446+
// content-defined chunking algorithm and returns a list of the chunk digests
447+
// in the order in which the chunks have to be concatenated to assemble the
448+
// requested blob.
449+
//
450+
// A client can expect the following guarantees from the server if a split
451+
// request is answered successfully:
452+
// 1. The blob chunks are stored in CAS.
453+
// 2. Concatenating the blob chunks in the order of the digest list returned
454+
// by the server results in the original blob.
455+
//
456+
// The usage of this API is optional for clients but it allows them to
457+
// download only the missing parts of a large blob instead of the entire blob
458+
// data, which in turn can considerably reduce download network traffic.
459+
//
460+
// Since the generated chunks are stored as blobs, they underlie the same
461+
// lifetimes as other blobs. However, their lifetime is extended if they are
462+
// part of the result of a split blob request.
463+
//
464+
// For the client, it is recommended to verify whether the digest of the blob
465+
// assembled by the fetched chunks results in the requested blob digest.
466+
//
467+
// If several clients use blob splitting, it is recommended that they request
468+
// the same splitting algorithm to benefit from each others chunking data. In
469+
// combination with blob splicing, an agreement about the chunking algorithm
470+
// is recommended since both client as well as server side can benefit from
471+
// each others chunking data.
472+
//
473+
// Errors:
474+
//
475+
// * `NOT_FOUND`: The requested blob is not present in the CAS.
476+
// * `RESOURCE_EXHAUSTED`: There is insufficient disk quota to store the blob
477+
// chunks.
478+
rpc SplitBlob(SplitBlobRequest) returns (SplitBlobResponse) {
479+
option (google.api.http) = { get: "/v2/{instance_name=**}/blobs/{blob_digest.hash}/{blob_digest.size_bytes}:splitBlob" };
480+
}
481+
482+
// Splice a blob from chunks.
483+
//
484+
// This is the complementary operation to the
485+
// [ContentAddressableStorage.SplitBlob][build.bazel.remote.execution.v2.ContentAddressableStorage.SplitBlob]
486+
// function to handle the splitted upload of large blobs to save upload
487+
// traffic.
488+
//
489+
// If a client needs to upload a large blob and is able to split a blob into
490+
// chunks locally according to some content-defined chunking algorithm, it can
491+
// first determine which parts of the blob are already available in the remote
492+
// CAS and upload the missing chunks, and then use this API to instruct the
493+
// server to splice the original blob from the remotely available blob chunks.
494+
//
495+
// In order to ensure data consistency of the CAS, the server will verify the
496+
// spliced result whether digest calculation results in the provided digest
497+
// from the request and will reject a splice request if this check fails.
498+
//
499+
// The usage of this API is optional for clients but it allows them to upload
500+
// only the missing parts of a large blob instead of the entire blob data,
501+
// which in turn can considerably reduce upload network traffic.
502+
//
503+
// In order to split a blob into chunks, it is recommended for the client to
504+
// use one of the servers' advertised chunking algorithms by
505+
// [CacheCapabilities.supported_chunking_algorithms][build.bazel.remote.execution.v2.CacheCapabilities.supported_chunking_algorithms]
506+
// to benefit from each others chunking data. If several clients use blob
507+
// splicing, it is recommended that they use the same splitting algorithm to
508+
// split their blobs into chunk.
509+
//
510+
// Errors:
511+
//
512+
// * `NOT_FOUND`: At least one of the blob chunks is not present in the CAS.
513+
// * `RESOURCE_EXHAUSTED`: There is insufficient disk quota to store the
514+
// spliced blob.
515+
// * `INVALID_ARGUMENT`: The digest of the spliced blob is different from the
516+
// provided expected digest.
517+
rpc SpliceBlob(SpliceBlobRequest) returns (SpliceBlobResponse) {
518+
option (google.api.http) = { post: "/v2/{instance_name=**}/blobs:spliceBlob" body: "*" };
519+
}
433520
}
434521

435522
// The Capabilities service may be used by remote execution clients to query
@@ -1778,6 +1865,59 @@ message GetTreeResponse {
17781865
string next_page_token = 2;
17791866
}
17801867

1868+
// A request message for
1869+
// [ContentAddressableStorage.SplitBlob][build.bazel.remote.execution.v2.ContentAddressableStorage.SplitBlob].
1870+
message SplitBlobRequest {
1871+
// The instance of the execution system to operate against. A server may
1872+
// support multiple instances of the execution system (with their own workers,
1873+
// storage, caches, etc.). The server MAY require use of this field to select
1874+
// between them in an implementation-defined fashion, otherwise it can be
1875+
// omitted.
1876+
string instance_name = 1;
1877+
1878+
// The digest of the blob to be splitted.
1879+
Digest blob_digest = 2;
1880+
1881+
// The chunking algorithm to be used. Must be IDENTITY (no chunking) or one of
1882+
// the algorithms advertised by the
1883+
// [CacheCapabilities.supported_chunking_algorithms][build.bazel.remote.execution.v2.CacheCapabilities.supported_chunking_algorithms]
1884+
// field.
1885+
ChunkingAlgorithm.Value chunking_algorithm = 3;
1886+
}
1887+
1888+
// A response message for
1889+
// [ContentAddressableStorage.SplitBlob][build.bazel.remote.execution.v2.ContentAddressableStorage.SplitBlob].
1890+
message SplitBlobResponse {
1891+
// The ordered list of digests of the chunks into which the blob was splitted.
1892+
// The original blob is assembled by concatenating the chunk data according to
1893+
// the order of the digests given by this list.
1894+
repeated Digest chunk_digests = 1;
1895+
}
1896+
1897+
// A request message for
1898+
// [ContentAddressableStorage.SpliceBlob][build.bazel.remote.execution.v2.ContentAddressableStorage.SpliceBlob].
1899+
message SpliceBlobRequest {
1900+
// The instance of the execution system to operate against. A server may
1901+
// support multiple instances of the execution system (with their own workers,
1902+
// storage, caches, etc.). The server MAY require use of this field to select
1903+
// between them in an implementation-defined fashion, otherwise it can be
1904+
// omitted.
1905+
string instance_name = 1;
1906+
1907+
// Expected digest of the spliced blob.
1908+
Digest blob_digest = 2;
1909+
1910+
// The ordered list of digests of the chunks which need to be concatenated to
1911+
// assemble the original blob.
1912+
repeated Digest chunk_digests = 3;
1913+
}
1914+
1915+
// A response message for
1916+
// [ContentAddressableStorage.SpliceBlob][build.bazel.remote.execution.v2.ContentAddressableStorage.SpliceBlob].
1917+
message SpliceBlobResponse {
1918+
// Empty for now, but might need to be extended in future.
1919+
}
1920+
17811921
// A request message for
17821922
// [Capabilities.GetCapabilities][build.bazel.remote.execution.v2.Capabilities.GetCapabilities].
17831923
message GetCapabilitiesRequest {
@@ -1964,6 +2104,36 @@ message Compressor {
19642104
}
19652105
}
19662106

2107+
// Content-defined chunking algorithms used for splitting blobs into chunks.
2108+
message ChunkingAlgorithm {
2109+
enum Value {
2110+
// No chunking. Servers MUST always support this, and do not need to
2111+
// advertise it.
2112+
IDENTITY = 0;
2113+
2114+
// Content-defined chunking algorithm based on Rabin fingerprinting. Details
2115+
// about the implementation can be found in algorithm 3 (RabinCDC8KB) of
2116+
// https://ieeexplore.ieee.org/document/9055082. This algorithm has the
2117+
// following properties:
2118+
// - minimum chunk size: 2 KB
2119+
// - maximum chunk size: 64 KB
2120+
// - average chunk size: 8 KB
2121+
RABINCDC_8KB = 1;
2122+
2123+
// Content-defined chunking algorithm based on Gear hashing. Details about
2124+
// the implementation can be found in algorithm 2 (FastCDC8KB) of
2125+
// https://ieeexplore.ieee.org/document/9055082. This algorithm has the
2126+
// following properties:
2127+
// - minimum chunk size: 2 KB
2128+
// - maximum chunk size: 64 KB
2129+
// - average chunk size: 8 KB
2130+
// The 256 64-bit random numbers in the Gear table are created with the
2131+
// Mersenne Twister pseudo-random generator for 64-bit numbers with a state
2132+
// size of 19937 bits and a seed of 0.
2133+
FASTCDC_MT0_8KB = 2;
2134+
}
2135+
}
2136+
19672137
// Capabilities of the remote cache system.
19682138
message CacheCapabilities {
19692139
// All the digest functions supported by the remote cache.
@@ -1997,6 +2167,11 @@ message CacheCapabilities {
19972167
// [BatchUpdateBlobs][build.bazel.remote.execution.v2.ContentAddressableStorage.BatchUpdateBlobs]
19982168
// requests.
19992169
repeated Compressor.Value supported_batch_update_compressors = 7;
2170+
2171+
// All the chunking algorithms supported by the remote cache. Remote cache may
2172+
// support multiple chunking algorithms simultaneously. Servers MUST support
2173+
// IDENTITY (no chunking), even if it is not listed here.
2174+
repeated ChunkingAlgorithm.Value supported_chunking_algorithms = 8;
20002175
}
20012176

20022177
// Capabilities of the remote execution system.

0 commit comments

Comments
 (0)