RDK Resources
[*RDK Preferred*]
Code Management Facility
RDK Forums
[RDK Conferences]
RDK Support
Archives
Papers & Presentations Archive
Expand | ||||
---|---|---|---|---|
| ||||
|
[warning
Warning | ||
---|---|---|
| ||
Draft page for review and comment prior to publication. |
...
Downloadable Application Containers (DAC) is a pan-RDK initiative to develop a container solution that allows binary applications to be downloaded and run on STBs without modification. DAC is initially targeting RDK-V set-top boxes but the aim is to extend DAC to suitable RDK-B devices in due course. DAC is a co-operation effort between Sky, Liberty Global, Metrological and Consult Red.
DAC builds upon open-source specifications, components and tools from the Open Container Initiative (OCI). In particular, it uses crun
, a fast, small, OCI-runtime compliant engine that can configure and launch applications containers. The first RDK contributions to be released are Dobby
and OCIContainers
. These are the first two components of an end-to-end system that operators can use to adapt, manage and control containers to be suitable for their platform. Other components within the end-to-end system that are currently in development include RDK-V SDK extensions that allow DAC Apps to be: built, published, and then adapted before loading and running on RDK-V STBs. The work for this is co-ordinated by RDK DAC Special interest Group (RDKDACSIG)
an RDK initiative and cooperation effort between Comcast, Sky, Liberty Global, RDK-M and Consult Red. The work for this is co-ordinated by RDK DAC Special Interest Group (RDKDACSIG)
The DAC end goal is to provide a solution that allows app developers to write and package also native (c/c++) Applications in a way that makes the application binary exchangeable and can be run across all RDK boxes without (compile time) modifications.
This group initially focused on RDK-V, the solution described on this wiki page is focused on RDK-V but many of the principles/same components are also suitable for RDK-B.
The diagram above provides a high level overview of the DAC solution.
It is divided into 4 functional areas which we describe futher in sections below:
This video, made by ConsultRed at TechSummit2024, nicely explains this bigger picture: dac_overview_rdk_tech_summit_2024_1080p.mp4
Multimedia name dac_overview_rdk_tech_summit_2024_1080p.mp4
Below is a more detailed diagram with technical components inside the RDK6.1 SW on STB and DAC reference cloud
draw.io Diagram | ||||||||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
3 functional area's are further discussed in sections below
A DAC SDK is created to make it easier for app developers to build oci, dac compliant containerized applications. The SDK is public, fully open source and uses standard Yocto 3.1 Dunfel Poky cross-compilation / build environment. It does NOT require you as an App developer to have any SoC or OEM specific SW/SDK for being able to compile and package your AppContainer, it is SOC/OEM agnostic. The SDK allows developers to cross compile their application sources, automatically add appropriate dependent libraries in AppContainer and produce oci, dobby, dac compliant AppContainer image. This binary oci AppContainer image can then be uploaded to an OCI registry for sharing with possible consumers/distribution platforms. As App developer you can test and run your AppContainer on all RDK6.1 RDK-M Video Accelerators (when publishing the App to RDK-M VA DAC cloud instance)
The DAC APP SDK is available on https://github.com/rdkcentral/meta-dac-sdk and comes with many example AppContainers
For information on how to make DAC app and use DAC SDK we refer to HOWTO build App with DAC SDK, cloudpublish and run on VA#useDACSDKandbuildyourfirstnativeDACapp
OCI container registry: is a standard Cloud Native component that allows storage and exchange of AppContainers in the standardized oci image format. see oci "Image Format Specification" by Open Container Initiative.
To allow distribution of its application ( to allow it to be run on RDK STBs), the App developer must upload its AppContainer to an oci container registry.
The OCI registry is the binary App exchange point between App developer and consuming parties/operators for distribution of the App on their target platforms.
To run the oci compliant AppContainer on set of devices in production, an associated secure distribution solution/system must be in place.
This distribution solution is responsibility of the operator and not of the app developer.
It is typically also a distribution system / instance per operator/environment because:
RDK provides a reference distribution solution, see the "DAC cloud reference" as depicted in reference architecture diagram above. It consists out of several µservices. If you hover over the specific µservice on the diagram you can see the link to its associated open source code.
RDK and RDK-M do not mandate use of a particular distribution solution towards the RDK operators.
The operators are free to choose, make/buy their secure distribution solution. Some operator choose to build further on system they already had, some operators are using their instance of the DAC cloud reference.
As long as the operator distribution system can take in the oci, dac compliant AppContainer, securely distribute and convert it to its OCI bundle equivalent ready for use on the operator specific CPE devices that support the dac contract for binary compatibility, it should be fine. In future we may want to validate this as part of firebolt certification of those devices.
RDK-M does operate an instance of the reference DAC cloud that serves an AppCatalog with DAC Appcontainers for all the RDK-M Video Accelerators
It is important staging, test and reference environment for App developers and RDK community, allowing App developers to publish their DAC native apps and test/run them on all the RDK Video Accelerators with a RDK6.1 software image (from 3 different SoC vendors).
Without such environment App developers cannot test and run their Application on real devices and that is obviously not acceptable.
If RDK-M organizes certification testing of a native Application, it will tested/certify against this environment
Operators will want to see the application successfully running and passing testing on this DAC Video accelerator environment before starting testing of the same app on their environment and boxes.
To be able to install and run the AppContainer on set of target devices/population following things are required:
The reference DAC cloud solution has been designed to fulfill those requirements.
In the architecture diagram you can find the bundle generator µservice that will upon RabbitMQ trigger automatically generate from an OCI image the appropriate OCI bundle for specific HWmodel (VA) and store it for downstream usage. It will use the appropriate config template for that associated HWmodel (see example HWmodel specific templates here ) for generation of the run config.json. The bundlegen code that the µservice is executing during this bundlegeneration of an app is located here.
In the architecture diagram you can find the bundle cryptor µservice that will upon generation of new bundle by bundlegeneration service, sign and encrypt the bundle. For more information on that we refer to DAC session at Technology Summit 2023#pane-uid-b13cbeba-87dc-4f0a-9932-5b7e00770663-1 and documentation on access restricted page DAC Security
That is what the reference DAC cloud system is doing. Once bundle for particular app_id/platform/compatible version is created, it will be available in cache of ASCS (Appstore Caching service) (which can serve as the origin for further caching on a CDN). If you request bundle for particular app_id/platform/compatibleversion towards ASCS and if it is not available in its cache it will automatically trigger real time creation of associated bundle in a flow managed by ASBS (appstore bundle service). ASBS will fetch appropriate metadata from ASMS and give instructions with right input parameters to respectively bundle-generator-service and bundle-cryptor-service via RabbitMQ to generate and secure the bundle and cache on ASCS.
The core components in any RDK software image to run the Appscontainer on CPE are "Dobby" and "crun". Dobby is an OCI Container manager aka container engine comparable to known desktop container engines dockerd and podman but then focused on embedded device, written in C++ and designed to have minimal footprint and performance impact.
Crun is an opensource container runtime alternative to "runc" equivalent but written in C, has smaller footprint but is also well supported by the industry (is for example default container runtime used by podman)
Dobby exposes API and is used to start, stop and monitor containers running on a STB, using the crun runtime underneath to interface with the kernel and start the containers.
When requested to start a new container, crun will process the associated container configuration (run config.json) and apply the associated container policies/settings.
The OCIContainer Thunder plugin allows for interfacing with Dobby using a JSON-RPC interface and Dobby can also be used to run other Thunder plugins in containers using the Dobby ProcessContainer backend. For more information on Dobby, see following wiki Dobby and for crun see the crun code repo.
Before being able to run the Appcontainer successfully on RDK CPE, other pieces are needed as well.
Some pieces may be operator-specific such as the UI to discover, store and run the apps but also not all operators will be using RDKshell as Appmanager and some will use alternative component to LISA for installation of apps. Also some may have system to distribute, store the apps onto the device without UI/user involved.
Operator device software must meet DAC contract for binary compatibility. Following firebolt native abstraction interfaces need to be in place: Rialto for audio, video & DRM and libGLESv2.so.2, libEGL.so.1, libwayland-egl.so.1 and libessos.so for Graphics. Also a Firebolt Server json-rpc server (which is "ripple" in RDK 6.1 reference environment) that supports current and upcoming Firebolt api version must be in place and pass certification tests (currently version 1.0 but we are working on and aim at 2.0 eg Lifecycle2 is required for native apps)
The RDK Video Accelerators running their RDK6.1 software image are the reference environment for RDK. They use the components as shown in the reference architecture diagram (VideoAccelerator UI, RDKshell with RialtoServermanager support, LISA, OCIcontainer rdkservice, OMI, ripple, westeros, dobby, crun. The Video Accelerator Resident App includes basic UI for working with DACapps. That UI allows you to browse and install DAC apps from the Appcatalog offered by ASMS in the DAC cloud instance setup for VA's. For demo video and more information on how to use that UI see video1 and video2.
That VA UI uses LISA API to get list of all installed apps at boot and for installing/uninstalling specific app.
For installation of a specific App, UI will ask for a particular app_id/app_version/HWplatform/compatibleFWversion combination the bundl_url to ASMS in DAC cloud and pass this bundle_url to LISA along with the install request. That URL will be https url pointing towards ASCS in DAC cloud with path towards actual (signed/encrypted) bundle tarball there. LISA will do the HTTPS request for that bundle tarbal. In regular case that bundle tarbal will be available in cache of ASCS or CDN in front and hence when LISA does the HTTPS request for that bundle tarbal, ASCS will be returned immediately with HTTP 200 OK. In case that bundle is not available in the cache yet (is the case when the bundle for that combination was not already pre-generated or requested by 1 other stb of same model), the ASCS will trigger ASBS in DAC cloud to automatically on the fly create the bundle for that combination (near Just in Time, eg 1-5 seconds order) and populate in the ASCS for download by LISA. In such "not in cache" case ASBS will return 202 iso 200 to LISA to indicate the near Just in Time generation and few seconds of patience. LISA has built-in, tunable by config, retry mechanism in case of such 202 so that it will also download the bundle in case of such near Just in time generation.
When successfully downloading the tarball LISA will unpack it in the appropriate directory locations it created for it, also create a separate persistent storage location for that app and add the app_id / version to its local database.
For more information on LISA we refer to LISA wiki https://wiki.rdkcentral.com/display/ASP/LISA
When the user wants to run one of the installed apps, the VA UI requests RDKshell, as the Appmanager, to start this app of type dac.native with as input the path to the associated bundle as stored by LISA.
RDKshell will create westeros wayland server instance/display socket for the app, create RialtoServer instance for the app (when needed), when Firebolt app interact with Firebolt Server (ripple) to get session_id for this app (is like security token, app needs when later connecting to Firebolt Server) and further delegate startContainer(bundle_path, westerosSocket) to OCIContainer thunder plugin which will in case of signed/encrypted container trigger the OMI component to verify signature and decrypt config.jwt in the crypted bundle and decrypt, verify and mount the rootfs of the Appcontainer rootfs using dm-verity & dm-crypt so it is ready to be consumed. Then the startContainer request is further delegated to Dobby which uses crun to execute it.
They will setup the container according to the associated run config.json, including bind mounts from STB host rootfs to Appcontainer rootfs of appropriate wayland and rialto server socket, as well as bind mounts of SoC specific Gfx library deps tree (that provide libGLESv2.so.2 and libEGL.so.1 )
When the dac.native app is a firebolt App, the specific Firebolt connection URL with session_id as security token, will be passed through the chain of RDKshell, OCI Container, dobby/crun and be available as an environment variable inside the Container. The app can then learn its value and use it when setting up connection with Firebolt Server.
Referring to the documentation HOWTO build App with DAC SDK, cloudpublish and run on VA
...
Used heavily in modern cloud software, containers are a standardised way to run the software in a self-contained, isolated and secure environment. Containers contain everything an application needs to run, including code, libraries and other dependencies. This means that there is no need to install specific libraries and dependencies on the host for each application. Unlike more traditional virtual machines, containers are lightweight and fast to create and destroy and don't have the overhead of virtualising an entire operating system. By sharing the OS kernel between containers, running applications inside containers adds very little performance or footprint overhead.
The most popular containerisation solution currently in use is Docker, although there are a number of other solutions such as LXC, Singularity and Podman. LXC containers have been available within RDK for a number of years, using a container generation tool at build time to define and create the container configurations but for use within monolithic STB software image, not for device agnostic building and downloading separately from monolithic image what DAC is achieving.
Image source: https://kubernetes.io/docs/concepts/overview/what-is-kubernetes/
There are a number of advantages of using containers to run applications in RDK deployments, including:
As part of the DAC initiative, containers are used to reduce the difficulty of developing native applications that can be run on many different RDK-V devices from different operators by creating an isolated environment for each application. This means the app becomes platform agnostic and can be run on devices the develop may not have physical access to.
From the Open Container Initiative (OCI) website (https://opencontainers.org/):
...
Formed in 2015 by Docker and other companies in the container industry and now part of the Linux Foundation, OCI define a number of specifications that allow developers to define containers. These specifications are followed by almost all major containerisation platforms.
OCI define both a runtime specification and an image specification. The Runtime Specification outlines how to run a “filesystem bundle” that is unpacked on disk. The OCI image is used for packaging containers in a platform-agnostic way that can be easily distributed. At a high-level, an OCI implementation would download an OCI Image then unpack that image into an OCI Runtime filesystem bundle. At this point the OCI Runtime Bundle would be run by an OCI Runtime.
draw.io Diagram | ||||||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
An OCI runtime is a CLI tool that allows for spawning and running containers according to the OCI specification. There are two main OCI runtimes in production use:
Repo: https://github.com/opencontainers/runc/
Runc is the reference implementation of an OCI runtime and is developed directly by the OCI group. This is the runtime used by Docker, Kubernetes and others. However, being written in Go, it is less suitable for embedded STB environments due to a relatively large footprint.
Repo: https://github.com/containers/crun
Crun is a different implementation of an OCI runtime, this time written in C and optimised for performance and a low memory footprint. It is developed at RedHat and is currently in use by Podman in Fedora, and is planned to be used in a main RHEL release soon. This is the runtime supported by Dobby and will be used in RDK.
Dobby is a container management tool, originally developed by Sky and open-sourced to the RDK community. Dobby is a daemon-based program that sits above the crun runtime, providing a more user-friendly experience and enhancing the base functionality provided by crun.
Dobby is designed to be used for starting, stopping and monitoring all containerised applications on the STB. To start a container with Dobby, simply provide Dobby with a path to an OCI bundle and a container ID, and Dobby will handle the rest.
Written in C++ and with a plugin-based architecture, Dobby is highly expandable and customisable, offering the ability to run custom code at various stages in the container lifecycle to add additional functionality. Out of the box, Dobby provides a number of plugins ready for use to provide commonly needed functionality including:
Plugins are simply C++ code written against the plugin interface, allowing for operators to easily add additional functionality, or modify existing functionality easily.
The core component of Dobby is the DobbyDaemon. This should be started at STB boot and then listens over dbus for commands. Dobby ships with a systemd unit file called sky-dobby.service, which allows Dobby to be started/stopped by systemd. DobbyDaemon logs are sent to journald.
Once the daemon is running, there are a number of ways to interact with Dobby
For debug builds of Dobby, the DobbyTool is installed on the STB. This is a simple command line app that communicates with Dobby directly over dbus to issue common commands. DobbyTool offers the following commands:
Code Block |
---|
vagrant@dobby-vagrant:~/srcDobby/build$ DobbyTool help
quit quit
help help [command]
shutdown shutdown
start start [options...] <id> <specfile/bundlepath> [command]
stop stop <id> [options...]
pause pause <id>
resume resume <id>
exec exec [options...] <id> <command>
list list
info info <id>
dumpspec dumpspec <id> [options...]
bundle bundle <id> <specfile> [options...]
set-dbus set-dbus <private>|<public> <address>
|
Some commands offered by DobbyTool (mainly dumpspec) are to allow compatibility with the legacy Dobby Spec input format (instead of OCI bundles), which are deprecated and maintained only for backwards compatibility with existing deployments.
To control Dobby using a JSON-RPC API, then use the OCIContainer Thunder NanoService. The source code for OCIContainer can be found here TODO:: UPDATE THIS ONCE OCICONTAINER IS MERGED: https://github.com/rdkcentral/rdkservices/pull/116.
OCIContainer code contains a README.md file with detailed documentation on its usage, and includes a test script (test/ociContainerTest.sh) that can be used to ensure Dobby/OCIContainer is working as expected.
To run an application as a Dobby container, the application must be packed as an OCI bundle. This bundle can come from many different sources - it could be manually created or generated from an OCI image.
For demonstration purposes, a sample bundle is attached to this page that creates a simple container. This container runs the sleep command to sleep for 30 seconds. It also configured Dobby's logging to log the container stdout/err to a file in /tmp/container.log
TODO:: Attach sleepy container bundle
For more information on how to create OCI bundles from scratch, see the following pages:
Dobby source code lives on the RDKCentral GitHub here: https://github.com/rdkcentral/Dobby and is Apache 2.0 licenced.
To develop Dobby, it is recommended to use Ubuntu 16.04. You cannot develop/test Dobby inside a Docker container as you cannot run Dobby containers inside a Docker container. TODO:: ADD LINK TO VAGRANT VM IF/WHEN WE MAKE IT PUBLIC
If you would like to contribute code to this project you can do so through GitHub by forking the repository and sending a pull request. Before RDK accepts your code into the project you must sign the RDK Contributor License Agreement (CLA).