CloudUUID

CloudUUID is a distributed unique ID generator grpc service backed by Google cloud spanner service. it supprot three kinds of algrithm: sequential, snowflake and cache snowflake

Squential ID

overview:

Each sequential id belongs to an unique id space, the space can be identified by the combination of the domain and tag. such as: (domain)order + (tag)prod will be the one unique Id space.

• Globally unique
• Scale to multiple workers
• 64 bits number(long)
• Sequential within one worker
• Preallocation and double buffered to achieve high qps

gprc function:

rpc createSequentiallIdSpace (CreateSequentialIdSpaceReq) returns (CommonResp) {}
rpc deleteSequentiallIdSpace (SequentialIdReq) returns (CommonResp) {}
rpc deleteAllIdSpaces (Empty) returns (CommonResp) {}
rpc getSequentialId(SequentialIdReq) returns (UuidResp) {}

SnowFlake ID

overview:

An unique id consists of worker node, timestamp and sequence within that timestamp. Usually, it is a 64 bits number(long), the timestamp is based on the true time of the spanner.

• Based on true time came from spanner system
• Id is generated strictly by true time
• the default bits of that three fields are as delta seconds, datacenter id, worker id, sequence

gprc function:

rpc getSnowFlakeId(Empty) returns (UuidResp) {}

Cached SnowFlake ID

overview:

it is basically the same as the snowflake model except that it takes buffered time.

• Use cached true time from spanner system
• Id is generated within the time diff tolerance
• double buffered to achieve high qps

gprc function:

rpc getCachedSnowFlakeId(Empty) returns (UuidResp) {}

---

Tech Features:

• Java with Spring boot framework
• Integration with stackdriver
• Integration with opencensus

installation

Precondition

you should have a gcp account, and you should create and download the service account json file.
[how to do this?](https://cloud.google.com/docs/authentication/getting-started). pls put the file 
into the class resource directory

1.create spanner

create the spanner instance [spnner](https://cloud.google.com/spanner/docs/quickstart-console).
and create the table:

CREATE TABLE uuidspace (
  NameSpace STRING(1024) NOT NULL,
  Tag STRING(1024) NOT NULL,
  MinId INT64 NOT NULL,
  MaxId INT64 NOT NULL,
  Description STRING(MAX),
  WorkerId STRING(1024) NOT NULL,
  Updated INT64 NOT NULL
) PRIMARY KEY (NameSpace, Tag)

2.config the application

Open application.properties
change the setting of :

gcp.project.id:xxx
gcp.service.account.file:xxx
spanner.instance:xxx
spanner.database:xx
spanner.database.table:uuidspace
uuid.range.stride:50000
uuid.space.buufer.base.level:10000
worker.id:1
datacenter.id:1
grpc.port: 8080

3.final.Start the service

java -jar clouduuid-1.0.0-SNAPSHOT.jar

Reference

• Leaf from meituan
• uid-gernartor from baidu

Limitation and known issues:

• bug of the imported plug "opencensus-exporter-trace-stackdriver"