diff options
Diffstat (limited to 'site/docs/documentation')
| -rw-r--r-- | site/docs/documentation/Input.md | 42 | ||||
| -rw-r--r-- | site/docs/documentation/Output.md | 66 | ||||
| -rw-r--r-- | site/docs/documentation/_category_.json | 7 |
3 files changed, 115 insertions, 0 deletions
diff --git a/site/docs/documentation/Input.md b/site/docs/documentation/Input.md new file mode 100644 index 00000000..8ea89936 --- /dev/null +++ b/site/docs/documentation/Input.md @@ -0,0 +1,42 @@ + +OpenDC requires three files to run an experiment. First is the topology of the data center that will be simulated. +Second, is a meta trace providing an overview of the servers that need to be executed. Third is the trace describing the +computational demand of each job over time. + +### Topology +The topology of a datacenter is described by a csv file. Each row in the csv is a cluster +of in the data center. Below is an example of a topology file consisting of three clusters: + +| ClusterID | ClusterName | Cores | Speed | Memory | numberOfHosts | memoryCapacityPerHost | coreCountPerHost | +|-----------|-------------|-------|-------|--------|---------------|-----------------------|------------------| +| A01 | A01 | 32 | 3.2 | 2048 | 1 | 256 | 32 | +| B01 | B01 | 48 | 2.93 | 1256 | 6 | 64 | 8 | +| C01 | C01 | 32 | 3.2 | 2048 | 2 | 128 | 16 | + + +### Traces +OpenDC works with two types of traces that describe the servers that need to be run. Both traces have to be provided as +parquet files. + +#### Meta +The meta trace provides an overview of the servers: + +| Metric | Datatype | Unit | Summary | +|--------------|------------|----------|--------------------------------------------------| +| id | string | | The id of the server | +| start_time | datetime64 | datetime | The submission time of the server | +| stop_time | datetime64 | datetime | The finish time of the submission | +| cpu_count | int32 | count | The number of CPUs required to run this server | +| cpu_capacity | float64 | MHz | The amount of CPU required to run this server | +| mem_capacity | int64 | MB | The amount of memory required to run this server | + +#### Trace +The Trace file provides information about the computational demand of each server over time: + +| Metric | Datatype | Unit | Summary | +|-----------|------------|---------------|---------------------------------------------| +| id | string | | The id of the server | +| timestamp | datetime64 | datetime | The timestamp of the sample | +| duration | int64 | milli seconds | The duration since the last sample | +| cpu_count | int32 | count | The number of cpus required | +| cpu_usage | float64 | MHz | The amount of computational power required. | diff --git a/site/docs/documentation/Output.md b/site/docs/documentation/Output.md new file mode 100644 index 00000000..10236958 --- /dev/null +++ b/site/docs/documentation/Output.md @@ -0,0 +1,66 @@ + +Running OpenDC results in three output files. The first file ([Server](#server)) contains metrics related to the jobs being executed. +The second file ([Host](#host)) contains all metrics related to the hosts on which jobs can be executed. Finally, the third file ([Service](#service)) +contains metrics describing the overall performance. An experiment in OpenDC has + +### Server +The server output file, contains all metrics of related to the servers run. + +| Metric | Datatype | Unit | Summary | +|-----------------|----------|---------------|-------------------------------------------------------------------------------| +| timestamp | int | datetime | Timestamp of the sample | +| server_id | string | | The id of the server determined during runtime | +| server_name | string | | The name of the server provided by the Trace | +| host_id | string | | The id of the host on which the server is hosted or `null` if it has no host. | +| mem_capacity | int | Mb | | +| cpu_count | int | count | | +| cpu_limit | float | MHz | The capacity of the CPUs of Host on which the server is running. | +| cpu_time_active | int | seconds | The duration that a CPU was active in the server. | +| cpu_time_idle | int | seconds | The duration that a CPU was idle in the server. | +| cpu_time_steal | int | seconds | The duration that a vCPU wanted to run, but no capacity was available. | +| cpu_time_lost | int | seconds | The duration of CPU time that was lost due to interference. | +| uptime | int | milli seconds | The uptime of the host since last sample. | +| downtime | int | milli seconds | The downtime of the host since last sample. | +| provision_time | int | datetime | The time at which the server was enqueued for the scheduler. | +| boot_time | int | datetime | The time at which the server booted. | + +### Host +The host output file, contains all metrics of related to the host run. + +| Metric | DataType | Unit | Summary | +|-------------------|----------|---------------|-------------------------------------------------------------------------------------------------| +| timestamp | int | datetime | Timestamp of the sample | +| host_id | string | | The id of the host given by OpenDC | +| cpu_count | int | count | The number of available cpu cores | +| mem_capacity | int | Mb | The amount of available memory | +| guests_terminated | int | count | The number of guests that are in a terminated state. | +| guests_running | int | count | The number of guests that are in a running state. | +| guests_error | int | count | The number of guests that are in an error state. | +| guests_invalid | int | count | The number of guests that are in an unknown state. | +| cpu_limit | float | MHz | The capacity of the CPUs in the host. | +| cpu_usage | float | MHz | The usage of all CPUs in the host. | +| cpu_demand | float | MHz | The demand of all vCPUs of the guests | +| cpu_utilization | float | ratio | The CPU utilization of the host. This is calculated by dividing the cpu_usage, by the cpu_limit | +| cpu_time_active | int | seconds | The duration that a CPU was active in the host. | +| cpu_time_idle | int | seconds | The duration that a CPU was idle in the host. | +| cpu_time_steal | int | seconds | The duration that a vCPU wanted to run, but no capacity was available. | +| cpu_time_lost | int | seconds | The duration of CPU time that was lost due to interference. | +| power_draw | float | Watt | The current power draw of the host. | +| energy_usage | float | Joule (Ws) | he total energy consumption of the host since last sample. | +| uptime | int | milli seconds | The uptime of the host since last sample. | +| downtime | int | milli seconds | The downtime of the host since last sample. | +| boot_time | int | datetime | The timestamp at which the host booted. | + +### Service +The service output file, contains metrics providing an overview of the performance. + +| Metric | DataType | Unit | Summary | +|------------------|----------|----------|------------------------------------------------------------------------| +| timestamp | int | datetime | Timestamp of the sample | +| hosts_up | int | count | The number of hosts that are up at this instant. | +| hosts_down | int | count | The number of hosts that are down at this instant. | +| servers_pending | int | count | The number of servers that are pending to be scheduled. | +| servers_active | int | count | The number of servers that are currently active. | +| attempts_success | int | count | The scheduling attempts that were successful. | +| attempts_failure | int | count | The scheduling attempts that were unsuccessful due to client error. | +| attempts_error | int | count | The scheduling attempts that were unsuccessful due to scheduler error. | diff --git a/site/docs/documentation/_category_.json b/site/docs/documentation/_category_.json new file mode 100644 index 00000000..0776466b --- /dev/null +++ b/site/docs/documentation/_category_.json @@ -0,0 +1,7 @@ +{ + "label": "Documentation", + "position": 5, + "link": { + "type": "generated-index" + } +} |