Permalink
Branch: master
Find file Copy path
Fetching contributors…
Cannot retrieve contributors at this time
229 lines (175 sloc) 25.2 KB

5. Summarized description per Test-bed component

This chapter contains for each component an overview of what it is aimed at, who uses it and in which phase of the Trial it is used, in the form of a table per component. The table also contains a hyperlink to the detailed specification of this component.

5.1. Common Information Space (CIS) {#cis}

Common Information Space
Short description The Common Information Space (CIS) describes the concept of information exchange between tools that are connected to the Test-bed. Also, information provided by simulations are forwarded and distributed to the tools via the CIS, and vice versa, messages sent from the CIS are transmitted to the CSS.
Who will use it Directly:
- Developers
Indirectly
- Everyone using a solution connected to the test-bed
Main functions The CIS will account for:
- Connecting the tools to the test-bed via CIS adapters
- Distribution of messages
- Security regarding access rights (They have not been defined)
Functions it does not do The CIS will NOT account for:
- Validation of messages
Links with other components - Admin tool – to configure (parts of) the CIS
- AAR – re-uses messages sent via the CIS
- CIS-CSS Gateways – data exchange between CIS and CSS in both directions
- Validation Service – messages coming from the CSS to the CIS are syntax validated
(Technical) conditions Apache Kafka is needed
Reference to repository/details -

5.1.1. Extra notes

The CIS itself is visible only to developers, not end-users. However, end-users may configure the CIS (or parts of it) to a certain extend using the Admin tool described in paragraph 5.4.

5.2. Common Simulation Space (CSS) {#css}

Common Simulation Space
Short description The Common Simulation Space (CSS) describes the concept of information exchange between Simulators that are connected to the Test-bed. These Simulators jointly generate and maintain a simulation world needed for providing a ground truth of the fictive crisis, feed the Solutions with fictive data via the CIS – and vice versa – receive messages sent from the CIS, and the deliver to the Participants a good enough image of the fictive crisis for them to be assessed.
Who will use it Implementers, Solution Owner, Simulator Owner, Test-bed Component Developers
Main functions The CSS will connect the simulators to each other and link to the gateways to exchange data CIS-CSS (bidirectional)
Functions it does not do Deal with security: simulators don't have secrets. Semantic validation/translation of messages send over the CSS. Other message format support (like JSON or XML).
Links with other components Individual simulators (bidirectional, but note these are no Test-bed components); Admin tool for configuration of the CSS; CIS-CSS gateways; Trial Scenario Manager (bidirectional); Time Service (from Time Service to CSS); AAR to use the data for analysis (from CSS to After Action Review)
(Technical) conditions Open source; Standardized connectors via the CSS Adapters; Possible to run in a Docker environment; Based on Apache Kafka
Reference to repository/details -

5.2.1. Extra notes

Simulators all have their own data model of how they represent the simulated world. The CSS allows these simulators to agree on a communication form that the simulators understand to create and maintain a joint simulated world. Next to the CSS, there also is the Common Information Space (CIS), that is used to connect all the solution tools with the Test-bed and thus with each other. The design to not connect the simulators to the CIS directly is mainly to ensure the two spaces of simulated truth and perceived/communicated truth are kept separate inside the Test-bed. Like with a lot of emergency management processes, obtaining relevant information from the real world to base a decision on is either done by: • actually being at a specific location observing the current state, or • receiving and sending messages via all kinds of communication channels from persons or systems at a specific location (e.g. radio communication, sensor input, camera feeds). These ways of obtaining information gives a (shared) perceived truth to be used in further emergency management decision making. However, due to a wrong observation or miscommunication, the perceived truth can be different than the simulated truth. The simulators should only be concerned with maintaining the current state of the simulated truth (including entities and processes), and shouldn’t have to deal with the different kinds of communication types for solution tools and users to create the perceived/communicated truth. The Common Simulation Space allows simulators to only focus on maintaining the current state of the simulated world (i.e. the simulated truth of the incident and the world around it). In order to communicate state changes with other simulators inside the CSS, self-created communication messages are allowed inside this space. This is different than the messages being sent over the CIS, because the CIS is more aligned with current emergency management standards (like Common Alerting Protocol (CAP) messages, or Emergency Data Exchange Language (EDXL) messages). In order to direct the simulated world towards a desired scenario relevant for the trial, the Trail/Scenario manager should be able to send out messages to change the simulated world. This should be done via inject messages that all simulators can understand in order to execute the requested inject (e.g. start the breach, let the container explode).

5.3. CIS-CSS gateways {#cis-css-gateways}

CIS-CSS gateways
Short description The CSS-CIS gateways are the interface between the Common Simulations Space (CSS) and the Common Information Space (CIS). Data from the simulations is translated into data that can be understood by the tools connected to the CIS and vice-versa. Since they translate specific message types, there may be many of them.
Who will use it Developers
Main functions The CSS-CIS gateways will account for:
- Aggregating messages coming from the CSS
- Translating between CSS topics and CIS topics
Functions it does not do The CSS-CIS gateways will NOT account for: Semantic translation
Links with other components - CIS & CSS: the gateways are the bridge between them
Validation service: the validation service is a gateway service that validates messages going through the gateway
(Technical) conditions Must run in a Docker environment
Reference to repository/details -

5.4. Validation Service {#validation-service}

Validation Service
Short description The message validation service serves the purpose of validating the syntax of messages to match a certain standard. Only successfully validated messages are distributed further inside the system. It is an extra step with respect to the default validation that is carried out by all adapters, and mainly required when testing new solutions or during a dry-run.
Who will use it It is internally used by the adapters, and configured via the Admin tool
Main functions The validation service will account for:
- Checking if the syntax of the message fulfils a certain standard
- Listening to a specific TOPIC (e.g. validation-cap-topic). if it validates, the messages will be forwarded to the corresponding standard topic (e.g. cap-topic).
Functions it does not do The validation service will NOT account for: Semantic validation of messages
Links with other components - CIS-CSS gateways: the message validation is a gateway service
CIS: only messages that fulfil a certain standard are forwarded to the rest of the system
(Technical) conditions Use of Apache AVRO schemas to represent standards
Reference to repository/details -

5.5. Test-bed manager (Admin tool) {#admin-tool}

Test-bed manager (Admin tool)
Short description The admin tool can be used to configure parts of the Test-bed, e.g. which tools are connected to each other (via the topics). Also, it provides Test-bed information (i.e. existing topics) as well as status information (heartbeat, configuration, logging) of the connected tools. Furthermore, it is the main interface to the security services.
Who will use it Trial manager
Main functions The Admin tool will account for:
- An interface to configure the CSS-CIS gateways
- An interface to configure the adapters
- Loading the KAFKA topic configurations
- Providing a Watch dog
Functions it does not do The Admin tool will NOT account for:
- Setting up topics after initialization
- Modifying topics after initialization
Links with other components - CIS-CSS gateways and adapters: they are configured using the admin tool
(Technical) conditions Must run in a docker environment
Reference to repository/details GitHub repository

5.6. Scenario Manager {#sm}

Scenario Manager
Short description The scenario manager can be used to create scenarios (master event lists) that are injected, via the Test-bed, into the CSS or CIS. For example, it injects a message to start a flooding, to send out emails to participants, or to instruct a role-player to perform an act. Naturally, it can also control the time, and (re-)start/pause/stop a scenario.
Who will use it Trial manager, scenario writers
Main functions Publish messages, often intended for simulators, but can also be used to directly show a message inside a solution.
Functions it does not do Record messages
Links with other components Can be used to send messages to simulators, but also to solutions, via the adapter. It's strongest connection is with the time service in order to control the fictive scenario time.
(Technical) conditions Must run in a docker environment
Reference to repository/details GitHub repository

5.7. Time service

Time service
Short description The Time service is the single source of truth of the fictive time during a trial. It listens to the scenario manager in order to play/pause/stop a scenario, and it may speed up or slow down the simulation. Each adapter will subscribe to the time service, and offer its users the fictive time. When sending messages containing time information, each service should use this fictive time. In addition, it shares the server's time using the Network Time Protocol (NTP), so all services in the test-bed can use it to sync their clocks.
Who will use it It is an internal service
Main functions Publish the fictive scenario time, play/pause/stop/speed up/slow down the time
Functions it does not do Run the simulation time backwards or time-jumps, as this would corrupt simulation processes. (Note that reloading a previously saved moment during a Trial is possible and is a way to do a time-jump backwards from an end-user's perspective.)
Links with other components Scenario manager and all adapters
(Technical) conditions Must run in a Docker environment
Reference to repository/details GitHub repository

5.8. Observer Support Tool {#ost}

Observer Support Tool (OST)
Short description The aim of OST is to collect observations, inform observers about trial progress, and visualize collected data
Who will use it Trial Manager (person who configures the OST up front of a Trial, manages Observers and has overall control over the system). User, which can be an Observer or Participant
Main functions Allowing trial managers to setup observation questions and send messages to observers
- Assigning observers to specific observation tasks
- Allowing observers to enter observations
- Allowing trial managers to monitor observations, also in real–time
- Allowing participants to complete surveys
- Storing observations and surveys, and offers means to analyse the data
Functions it does not do OST does not provide sets of data about events and time. This tool waits for simulation phases from external system, publish events and generate questions based on these data.
Links with other components OST is linked with Test-bed, which is responsible for providing simulation phases – data about events and simulation time. Package of data is sent to OST Server, events and set of questions are generated. When the new event is displayed, OST Server notifies the user and set of questions changes.
(Technical) conditions Observer Support Tool (Mobile):
- Web browser
- Android (tablet and smartphone)
- iOS (tablet and smartphone)

Observer Management Tool (Desktop):
- Web browser
Reference to repository/details https://github.com/DRIVER-EU/ost or specifications

5.8.1. Extra notes

The Observer Support Tool’s aim is to collect observations, inform observers about trial progress and visualize collected data. There are different perspective to look at this tool. Main user, who uses the mobile version of a tool to send his observations is called Observer. From the other side there exists Trial Manager, he focuses on collected data and analysing it on desktop. Each of them has their own functionalities provided by OST. There are also these functionalities which are connected with non-functional requirements.

Observer Support Tool provides different views for each user. Observer sees name and description of a trial, events that have already happened and observation templates which he can fill in, whereas Trial Manager have displayed summary of all observations that have been sent, what is more he can even see summary of observations in time and messages he sent to Observers. Trial Manager is responsible for assigning role to the user which can be an Observer or Participant of a Trial.

First, user selects trial he is interested in. He sees its name, description and list with events that have already happened. User can send his Observation by answering questions that are connected with events – each new event is a trigger and can change the set of questions.

OST Server does not provide data about events, it is responsible for data exchange but not for preparing them. It receives data packages about events and simulation time from Test – bed and reacts on triggers. Events can be also sent directly to users by Trial Manager.

If events are sent from Trial Manager, OST Server publishes them both to Test-bed and to the user. Trial Manager not only manages the trial and user but also prepares environment to obtain information that are needed. He is responsible for projecting questions and question types. Contents of questions and number of them are optional, only the form is imposed.

When Test-bed sends package with data about events and time, OST Server notifies Trial Manager and users about new event. OST Server reacts on triggers and matches proper set of questions, which are sent to Trial Manger and then published to Observer. When Observer sends his Observations, Trial Manager collects obtained data and has displayed reports about them and if he needs it he can generate it in CSV.

Questions have different purposes, which are indicated in Observation Types. Great advantage of it is getting better criteria of comparison. Database with observations is more varied but it also connects categories, so analysing is more efficient. Correctly prepared questions and labels lead to most efficient results and better conclusion. Questions can have also different answer type such as slider, checkboxes, radio buttons and text field. With sending Observation user can also add some extra material such as additional description, voice record, picture or location. Each observation refers to participant and enables change of time.

5.9. After-Action-Review (AAR) {#aar}

After-Action-Review
Short description The After-Action Review (AAR) tool provides the possibility to collect data after a trial has finished and analyse it. Its main purpose is to facilitate the evaluation of the trialled solutions, and to help the participants determine how well they functioned. It collects messages (exchanged during trial), observation reports and takes screen-shots.
Who will use it Facilitator
Main functions The AAR tool will account for:
- Storing relevant data
- Reviewing a trial completely or parts of it
- Jump to specific point in time
Functions it does not do The AAR tool will NOT account for: Changing the course of the trial afterwards in any way
Links with other components CIS, CSS, Scenario Manager, Observer Support Tool: The AAR tool uses the messages exchanged inside the CIS, CSS, Scenario Manager, and the observations to provide the review capability.
(Technical) conditions The Security Service may have to grant the access to all secured topics to the AAR backend service to give the service the possibility to collect the data.
Reference to repository/details GitHub repository

5.10. Security Services

Security Services
Short description The Security Services provide access control enforcement on the DRIVER+ Test-bed, as well as support functions for identity and access management.
Who will use it Directly: developers, test-bed administrators. Indirectly: everyone using a solution connected to the Test-bed.

| | Main functions | The Security Services will account for:
- Topic access policy enforcement; i.e. the Admin tool defines an access policy (set of access rules) per CIS topic, for the topics with confidentiality requirements, and delegates to the Security Services the enforcement of such policies in CIS.
- SSL client certificate management, for managing (mostly issuing) SSL client certificates of CIS adapters for each tool connected to the Test-bed; these certificates are required for test-bed-level client authentication. | | Functions it does not do | The Security Services will not account for:
- The SSL authentication on CIS middleware (Kafka); this will rely on Kafka existing features, only the trusted CA certificate used in Kafka configuration comes from Security Services. (Not to be confused with the subsequent authorization phase that will be handled by Security Services indeed, via Kafka extension.)
- Protection, escrow or recovery of secret/private keys. Tools will have the possibility to have the Security Services generate keypairs (with the certificates) for them, for Test-bed purposes only. However, the Security Services are not responsible for the protection, escrow or recovery of the generated private (or secret) keys in any way. If the certificate holder loses them, new ones – with a new certificate - will be generated.
- Attack detection and/or mitigation. | | Links with other components | - The Admin Tool consumes the Security Services for topic access policy configuration and enforcement in CIS.
- The CIS trusts the Security Services CA for client authentication.
- The CIS consumes the Security Services for topic access control. | | (Technical) conditions | - Apache Kafka is needed.
- Host OS should be Ubuntu 16.04 LTS 64-bit (or more recent), RAM >= 4GB, file-system ext4, disk space >= 40 GB, Gigabit LAN connection
- Strongly recommended (but not mandatory): OS/applications should have access to a TPM preferably; else at least to high-quality entropy for cryptographic pseudo-random number generation. | | Reference to repository/details | https://projectdriver.sharepoint.com/:p:/r/sites/DriverPlus/Documents%20partages/SP92%20-%20Testbed/WP923%20-%20Testbed%20infrastructure/Meetings/2018-02-12.14%20Vienna%20Development%20meeting%20February/Presentations/AccessControl-SecurityRoadmap-V3.pptx?d=w01f66b178f0143ad9eb3179fc3f8fc18&csf=1&e=VyUcKc

https://projectdriver.sharepoint.com/:w:/r/sites/DriverPlus/Documents%20partages/SP92%20-%20Testbed/WP923%20-%20Testbed%20infrastructure/Meetings/2018-02-12.14%20Vienna%20Development%20meeting%20February/DRIVER+%20F2F%20Minutes%20-%20Vienna%20-%202018-02-12.14.docx?d=w17d0df9098714babbeb746c0015a3a67&csf=1&e=UsdOiV |

5.11. Play service {#play}

Play service
Short description The Play service acts as a mini-scenario editor: it can either publish one message, or play a sequence of (timed) messages. Besides being useful for debugging, where a developer can replay a recorder scenario, it is also useful for testing solutions standalone, where the play service can play a simple scenario so the solution can shine. The recorded messages can be obtained using Landoop's Kafka Topics UI, which allows you to download all messages in a topic to a single JSON file with key-value messages.
Who will use it Developers; Participants; Solution owners/evaluators
Main functions Publish a single recorded message, or a sequence of messages, to the test-bed
Functions it does not do Create messages from scrap, or record messages
Links with other components It is connected to the test-bed via an adapter, and it is related to the message injector
(Technical) conditions Should run in a Docker environment, can get recorded messages from a mounted Docker volume
Reference to repository/details GitHub repository

Related to this service is the open source csCOP (Common Operational Picture) tool, which can be used as a COP during a trial, but which can also be used to test the messages published by simulators and solutions.

5.12. Message Injector

Message Injector
Short description The message injector is used for debugging purposes, similar to Postman, and it is used to inject messages manually into the Test-bed in order to generate a certain response.
Who will use it Developers
Main functions The main functionality is to inject (prepared or on the fly) messages into the test-bed for testing purposes
Functions it does not do -
Links with other components It will use one of the existing adapters, most likely, the Java one, to send its messages to the test-bed
(Technical) conditions Must run in a Docker environment
Reference to repository/details -

5.13. Data services

Data services
Short description The Data services are not a single service, but a group of non-essential, but very practical services to complement the Test-bed. Their main purpose is to share data to enrich the Trial, such as map data, height data, census layers, weather information, et cetera.
Who will use it Trial staff
Main functions Share map data, height data, vector data
Functions it does not do Create or edit this data
Links with other components Can be used by solutions as well as simulators as a data backend
(Technical) conditions Must run in a Docker environment
Reference to repository/details WMS service, MBtiles service

Related to these data services is the AVRO schemas repository, which contains all the schema's that have been used (so far).

5.14. Docker-composer

Docker-composer
Short description All of the Test-bed's core functionality run inside a Docker environment (virtual machines). The Docker-composer website allows you to select the test-bed components you actually need, and it creates a dedicated Docker-compose file for you. This file can be easily run (docker-compose up -d), and the test-bed is started using a single command, linking together services and data.
Who will use it End users wishing to try out a solution, Developers and Sysops to setup the test-bed
Main functions Tie many different Docker images together
Functions it does not do Create a Docker image for you
Links with other components All Dockerized services, including the Dockerized solutions available in the Portfolio of Solutions
(Technical) conditions Must run in a Docker environment
Reference to repository/details Website, GitHub repository,