Hi Cassandra database experts,
Can you help us estimate the Cassandra processing power based on below hardware and cluster configuration ?
They are current price-performance sweet spot hardware.
Data Model: Only one table
create table_1
(
key_1 varchar(30),
key_2 varchar(30),
key_3 varchar(30),
key_3 varchar(30),
col_1 varchar(30),
col_2 varchar(30),
col_3 varchar(500),
primary key (key_1)
);
Besides primary key index, there are 3 indexes, are respective on column key_2, key_3 and Key_4.
There are 60 million rows.
Average row length 500 bytes.
Memory: 16GB to 64GB
CPU: 8-core CPU processors
Disk:
- SSD (solid state drives) : Size ?
- SATA spinning disks : Size ?
- Data Model 1: One base table with 3 indexes
- Data Model 2: One base table and 3 Materialized View tables.
- Data distribution and replication
- Strategy 1: one data center, 3 nodes, replication_factor = 3. Write Consistency Levels = 2
- Strategy 2: two data centers, 3 nodes on each data center,
The final matrix will be looked like this:
Read/Write operation pattern
|
Max Throughput
|
Response time
| |||
99% Reads
|
99.99% Reads
|
99% Writes
|
99.99% Write
| ||
100% read
|
? reads/second
|
< ? seconds
|
< ? seconds
| ||
99% read, 1% write
|
? reads/second,
? writes/second
|
< ? seconds
|
< ? seconds
|
< ? seconds
|
< ? seconds
|
90% read, 10% write
|
? reads/second
? writes/second
|
< ? seconds
|
< ? seconds
|
< ? seconds
|
< ? seconds
|
50% read, 50% write
|
? reads/second
? writes/second
|
< ? seconds
|
< ? seconds
|
< ? seconds
|
Less than ? seconds
|
Disk storage size : ? GB.
Please help to fill the green text ? with estimated numbers.
If you could tell us how did you calcuate these number, it will be much better.
Thanks,
Charlie 木匠 | Database Architect Developer
No comments:
Post a Comment