@@ -9,12 +9,12 @@ brought into being.
9
9
The Snowflake-SI-vA algorithm uses 53-bits integer as the output of UUID
10
10
generation, its bits-mapping shows like following:
11
11
12
- | H | 52 ~ 10 | 9 ~ 0 | L |
13
- | :-:| :-------------------------------------------: | : ----------:| :-:|
14
- | | 1111111111111111111111111111111111111111111 | 1111111111 | |
15
- | | UIN | MID | |
12
+ | H | 52 ~ 43 | 42 ~ 0 | L |
13
+ | :-:| :----------: | : -------------------------------------------:| :-:|
14
+ | | 1111111111 | 1111111111111111111111111111111111111111111 | |
15
+ | | MID | UIN | |
16
16
17
- As above graph shows, the high 43 bits are the UIN(UUID Index Number). It's the
17
+ As above graph shows, the low 43 bits are the UIN(UUID Index Number). It's the
18
18
sum of the cursor of an incremental sequence and the current Unix timestamp in
19
19
milliseconds. It means the genrated UUID is still an incremental sequence.
20
20
@@ -23,7 +23,7 @@ milliseconds. It means the genrated UUID is still an incremental sequence.
23
23
> incremental sequence and timestamp, there is 2-bits as the extra reserved
24
24
> zone.
25
25
26
- The low 10-bits is the ID of machine, for distributed services. So that it
26
+ The high 10-bits is the ID of machine, for distributed services. So that it
27
27
will not generate duplicated UUID between multi-instances. Different to the
28
28
Snowflake-SI, Snowflake-SI-vA use a fixed 10-bits for MID, and then there could
29
29
always be 1024 instances at most.
0 commit comments