The document discusses a Meetup presentation by Eric Evans on time-series data storage using Cassandra and its capabilities for high-throughput, horizontally scalable storage and retrieval of grouped metrics. It highlights the limitations of the existing RRDTool system and presents Newts as a solution with optimized writing and retrieval of sample data, structured as a key-value format. Key features include the aggregation of measurements at query time and a sample structure for data organization.

1. Time Series Storage
Cassandra London Meetup
April 7, 2014
Eric Evans
eevans@opennms.org
@jericevans

2. Open

3. Open

4. Network
Management
System

5. OpenNMS: What It Is
● Network Management System
○ Discovery and Provisioning
○ Service monitoring
○ Data collection
○ Event management and notifications
● Java, open source, GPLv3
● Since 1999

6. Graph All The Things

7. RRDTool
● Round robin database
● First released 1999
● Time-series storage
● File-based
● Constant-size
● Automatic, amortized aggregation

8. Consider
● 2 IOPs per update (read-update-write)
● 1 RRD per data source (storeByGroup=false)
● 100,000s of data sources, 1,000s IOPS
● 1,000,000s of data sources, 10,000s IOPS
● 15,000 RPM SAS drive, ~175-200 IOPS

9. Also
● Not everything is a graph
● Inflexible
● Incremental backups impractical
● ...

10. Observation #1
We collect and write a great deal; We read
(graph) relatively little.
We are optimized for reading everything,
always.

11. Observation #2
Samples are naturally collected, and graphed
together in groups.
Grouping samples that are accessed together
is an easy optimization.

12. Project: Newts
Goals:
● Stand-alone time-series data store
● High-throughput
● Horizontally scalable
● Grouped metric storage/retrieval
● Late-aggregating

13. Cassandra
Why:
● Write-optimized
● Sorted
● Horizontally scalable (linear)

14. Gist
● Samples stored as-is.
● Samples can be retrieved as-is.
● Measurements are aggregations calculated
from samples (at time of query).

15. Sample
{
“resource” : “london”,
“timestamp” : 1396289065,
“name” : “meanTemp”,
“type” : “GAUGE”,
“value” : 17.2,
“attributes” : { “units”: “celsius” }
}

16. Samples
CREATE TABLE newts.samples (
resource text,
collected_at timestamp,
metric_name text,
metric_type text,
value blob,
attributes map<text, text>,
PRIMARY KEY(resource, collected_at, metric_name)
);

17. Samples
resource | collected_at | metric_name | value
---------+---------------------+--------------+-----------
london | 2014-03-31 18:04:25 | dewPoint | 0xc01a0000
london | 2014-03-31 18:04:25 | maxTemp | 0x40280000
london | 2014-03-31 18:04:25 | maxWindGust | 0x7ff80000
london | 2014-03-31 18:04:25 | maxWindSpeed | 0x40180000
london | 2014-03-31 18:04:25 | meanTemp | 0xbfe00000

18. Behind the scenes...
london (2014-03-31 18:04:25, dewPoint):
0xc01a0000
(2014-03-31 18:04:25, maxTemp):
0x40280000
...
Ascending Order

19. http://github.com/OpenNMS/newts