**ReCodEx** is designed to be very modular. **WebApp** is considered to be frontend part of whole system while all other parts/services are considered to be backend part. **WebApp** contains almost all logic of the app including _user management and authentication_, _storing and versioning files_ (through connection with **Fileserver**), _counting and assigning points_ to users etc. **WebApp** is connected with **Broker**, **Fileserver** and partially with **Monitor**. **Broker** is essential part of whole architecture and can be marked single point of failure.
**ReCodEx** is designed to be very modular. **WebApp** is considered to be frontend part of whole system while all other parts/services are considered to be backend part. **WebApp** contains almost all logic of the app including _user management and authentication_, _storing and versioning files_ (through connection with **Fileserver**), _counting and assigning points_ to users etc. **WebApp** is connected with **Broker**, **Fileserver** and partially with **Monitor**. **Broker** is essential part of whole architecture and can be marked as single point of failure.
Almost whole communication goes through **Broker** and `ZeroMQ` messaging middleware. When **WebApp** wants to execute submission then all datas are handed over to **Worker** through **Broker**, similar situation is with progress state which start in **Worker** goes through **Broker** then pass **Monitor** and end up in **WebApp**. Only part of communication, which does not include **Broker**, is communication with **Fileserver** which is realized through `HTTP` commmunication. This communication can be initiated by **WebApp** or by **Worker**, other services have no access to **Fileserver**. Summarize of above stated can be found in overall architecture image (green arrows = `HTTP` communication, red arrows = communication through `ZeroMQ`).
Almost whole communication goes through **Broker** and `ZeroMQ` messaging middleware. When **WebApp** wants to execute submission then all datas are handed over to **Worker** through **Broker**, similar situation is with progress state which start in **Worker** goes through **Broker** then pass **Monitor** and end up in **WebApp**. Only part of communication, which does not include **Broker**, is communication with **Fileserver** which is realized through `HTTP` commmunication. This communication can be initiated by **WebApp** or by **Worker**, other services have no access to **Fileserver**. Summarize of above stated can be found in overall architecture image (green arrows = `HTTP` communication, red arrows = communication through `ZeroMQ`).
@ -16,27 +16,23 @@ Almost whole communication goes through **Broker** and `ZeroMQ` messaging middle
**Worker's** main role is securely _compile_, _run_ and _evaluate_ given submit against model solutions provided by author of each task. It is logicaly divided into three objects:
**Worker's** main role is securely execute given submission and possibly _evaluate_ results against model solutions provided by submitter. **Worker** is logicaly divided into two parts:
- **Message Frontend** communicates with **WebApp** using messaging queue [ZeroMQ](http://zeromq.org/). It receives new submits, operates the evaluation through **Work API** and reports progress back.
- **Listener** - listens and communicates with **Broker** through [ZeroMQ](http://zeromq.org/). It receives new jobs, communicates with **Evaluator** part and sends back results or progress.
- **Worker Core** can do all evaluating steps and is responsible for security of them. Sandbox [Isolate](https://github.com/ioi/isolate) is used.
- **Evaluator** - gets jobs to evaluate from **Listener** part, evaluate them (possibly in sandbox) and get to know to other part that evaluation ended. This part also communicates with **Fileserver**, downloads needed files and uploads detailed results.
- **File Server Frontend** provides access to files on **File Server** via
**File API**, where testing inputs and corresponding outputs for each task and
other required files are stored. It's possible to upload files, too.
The worker has to:
**Worker** after getting evaluation request has to:
- Download the archive containing the submission and an isoeval configuration
- Download the archive containing submitted source files and configuration file
file
- Download any supplementary files based on the configuration file, such as test
- Download any supplementary files based on the configuration file, such as test
inputs or helper programs (This can be done on demand, using a `fetch` command
inputs or helper programs (This is done on demand, using a `fetch` command
in the assignment configuration)
in the assignment configuration)
- Download the source codes of the student's submission
- Evaluate the submission accordingly to job configuration
- Evaluate the submission according to the assignment's configuration
- During evaluation progress states can be sent back to **Broker**
- Upload the results of the evaluation to the file server
- Upload the results of the evaluation to the **Fileserver**
- Notify the broker that the evaluation is finished
- Notify **Broker** that the evaluation finished
### Internal Worker architecture
### Internal Worker architecture
Picture below is overall internal architecture of worker, which shows its defined classes with private variables and public functions. Vector version of this picture is available [here](https://github.com/ReCodEx/GlobalWiki/raw/master/images/Worker_Internal_Architecture.pdf).
Picture below is overall internal architecture of worker which shows its defined classes with private variables and public functions. Vector version of this picture is available [here](https://github.com/ReCodEx/GlobalWiki/raw/master/images/Worker_Internal_Architecture.pdf).
Martin Polanka
8 years ago