Since RCE 10, RCE provides three possibilities to connect your RCE instance to other RCE instances and to use the user-integrated tools and components published on those instances: The RCE network connections, SSH Uplink connections and SSH Remote Access connections. RCE connections are meant to be used only in a trusted network (e.g. your institution's internal network). The RCE network traffic is currently not encrypted . This means that it is not secure to expose RCE server ports to untrusted networks like the internet. In the case that it is not possible or not secure to use RCE connections, SSH connections provide a more secure alternative.
As the new Uplink connections do not yet support all features of the former SSH connections (the publishing of workflows is not possible by Uplink connections), we decided to keep both types of connections in the current release. Thus, in the network view there are now 3 types of connections: the standard RCE connections (meant to be used in secure internal networks), the old "SSH Remote Access Connections" and the new "Uplink Connections".
The following table compares the three connection types:
Table 3.13. Connection types - feature matrix
| Connection type | RCE connections ("internal network") | SSH Remote Access connections | SSH Uplink connections |
|---|---|---|---|
| Publishing built-in tools (e.g. Joiner, Parametric Study, ...) | yes | no | no |
| Publishing user-integrated tools | yes | yes | yes |
| Publishing workflows as tools | no * | yes | no * |
| Symmetric/bidirectional tool publishing | yes | no | yes |
| Accessing remote workflow status and data management | yes | no | no |
| Remote system monitoring (CPU/RAM) | yes | yes | no ** |
| Login authorization (via password or SSH keyfile) | no | yes | yes |
| Suitable for insecure networks (e.g. internet) | no (!) | partially *** | yes (via relay) |
* = planned for RCE 11; ** = may be added in a future release; *** = connections
are encrypted, but require an open incoming network port for publishing tools - if possible,
use Uplink connections instead
RCE connections are meant to be used only in a trusted network (e.g. your institution's internal network). To build up a network of RCE instances, at least one of the instances has to be configured as a server (see the "Configuration" section or the sample configuration file "Relay server" for details).
On the client side, RCE network connections can be added in the "network" view by clicking "Add network connection" and entering the hostname and port of an RCE server instance. The connections are shown in the "RCE Network"->"Connections" subtree. They can also be edited, connected and disconnected in the network view. However, the changes made here are not saved in the configuration yet, i.e. they will be lost when RCE is closed or restarted. To permanently add connections, you can edit the configuration file (see the "Configuration" section for details).
In the "RCE Network" -> "Instances" subtree all RCE instances in the network are listed. When expanding the entry for an instance, you can see monitoring data like CPU or RAM usage for this instance, and the published components and tools of this instance (if it has any).
The published components and tools of the other instances in your network are also shown in the palette of the Workflow Editor. From there, you can use them in your workflows just like your local components and tools. When you start a workflow, in the "Execute Workflow" wizard there is an overview which component will be run on which RCE instance. If a component is available on several instances, you can choose here on which instance it should be run. In the same wizard, you can also choose another instance as the "Controller Target Instance", which means that the workflow execution will be controlled by this instance (see the section "Configuration Parameters" for more information). This can be useful when you start a long-running workflow where all components are run on remote instances and you do not want to keep your local computer connected all the time.
Uplink connections allow to use the "SSH relay" functionality. This means that it is possible to setup a single server as the "relay" for a project (and only this server needs to be reachable on an SSH port). All other RCE instances can connect to this server as clients via SSH Uplink Connections and publish their tools so that they can be used by other clients. (In contrast, with the former version of SSH connections every partner who wanted to publish tools needed to configure an SSH server).
The RCE instance that should be used as the relay has to be configured as an SSH server and provide at least one account with the role "uplink_client" or "remote_access_user"(see Section 2.2, “Configuration” or the sample configuration file "Uplink relay" for details). The recommended role is "uplink_client", which allows only access to Uplink connections and no access to an interactive SSH shell.
Note
When configuring an SSH account using a key file, both server and client have to run RCE
7.1 or newer. In RCE 10.0.0, only RSA Keys generated by the tool
puttygen using Windows-style line endings work. This is a known
issue with RCE 10.0.0 and will be fixed in an upcoming version of RCE.
When using Windows, the default settings of puttygen, which comes
bundled together with the popular SSH client putty, are sufficient.
When using Linux, you will have to install the tool puttygen.
Please refer to the documentation of your system for instructions on installing
that tool. After you have generated a key on Linux, you will have to convert it
to use Windows-style line endings. We recommend the tool todos for
this task. Both puttygen and todos are readily
available for most major distributions from the official package sources.
On the client side, Uplink connections can be added in the "network" view by clicking "Add Uplink Connection". In the following dialog, enter the hostname and port of an Uplink relay as well as the user name and the authentication data of an SSH account configured on this instance. Depending on the SSH account, you have to authenticate using a passphrase or by an RSA private key file. If your private key is protected by a passphrase, select the authentication type "Keyfile with passphrase protection", else select "Keyfile without passphrase protection". If several clients are using the same account on a relay, enter a different "client ID" on each of them.
If the instance should serve as a gateway (i.e. forward tools between the (external) Uplink network and a local network), set the "isGateway" parameter to "true".
The connections are shown in the "Uplink"->"Uplink Connections" subtree. They can also be edited, connected and disconnected in the "network" view. It is possible to store passphrases using the Eclipse Secure Storage Mechanism. However, the changes made here are not saved in the configuration yet, i.e. they will be lost when RCE is closed or restarted. To permanently add Uplink connections, you can edit the configuration file (see Section 2.2, “Configuration” for details). Sample configuration files are avaible as "Uplink Client" and "Uplink Gateway".
Configuring a gateway in non-GUI mode involves four steps:
-
Configure an SSH Uplink connection to the SSH relay server in the profile's
configuration.jsonfile. In this connection, make sure to set the "isGateway" parameter to "true" (without quotes). -
Configure a normal RCE server port for the internal network. This is the network port that clients in the local (internal) network can connect to with standard ("internal network") connections.
-
Using the file-based import feature (see section Section 2.2.4, “Importing authorization data without GUI access”), import the SSH password or the SSH keyfile passphrase for logging into the Uplink relay. (Please note that currently, the gateway must be (re)started after creating these import files to apply the changes.)
-
To allow the gateway to forward tools that are not public, but only published for specific authorization groups, the gateway must be a member of at least one matching group. Use the file-based import feature (see section Section 2.2.4, “Importing authorization data without GUI access”) to import any required group keys. (Please note that currently, the gateway must be (re)started after creating these import files to apply the changes.)
In order to make tools available for other clients, you have to publish them (for example using the "Component Publishing" view; see user guide for more information about pubishing/authorization groups). To make a tool available via an SSH relay, it has to be either in the "Public Access" group or in an authorization group which name starts with "external_". Tools in other authorization groups will only be shared in your local RCE network.
Note
Note: Tools that are available to a client via an Uplink connection are also available to RCE instances connected to that client in its local RCE network (if they possess the corresponding authorization group key and the "isGateway" option is set for the Uplink connection). Accordingly, tools published by those instances in the "Public Access" group or in an authorization group which name starts with "external_" will also be made available via the Uplink relay. Please not that this only works if the gateway itself also possesses the authorization group key.
Nodes connected via Uplink connections do not show up in the network view as nodes (same as Remote Access).
Imported tools show up in the Network view under the Node running the Uplink connection (also the same as Remote Access), and they are not yet marked or distinguishable from normal components.
Tools located on the RCE instance serving as relay are not published automatically. If you want to publish them, you have to add a connection to the relay from the same instance.
Uplink connections are an experimental feature in RCE 10.x and have some known limitations:
-
Connections are always encrypted as part of the SSH connection, but there is no additional "internal" encryption of tool input/output data yet (which is planned for future versions to protect users against untrustworthy relay servers).
-
The behavior on errors, disconnects, and server shutdowns is not fully implemented yet; this will be stabilized in RCE 11.
-
Custom tool icons are not yet transferred over Uplink connections.
The following figure gives an example of how a cross-organization network using Uplink connections could be structured:
The four project partners in the example all have an internal network of RCE instances which are connected by standard RCE connections. Uplink connections to a relay server are used to connect between the different partners. The relay server is located outside of the organizations networks, and only the relay server has to be reachable via SSH over the internet. Typically, for each organization one RCE instance (called SSH gateway) established an SSH connections to this relay server. All other instances in the institution’s internal network can be connected to it by standard RCE connections and still publish tools to the other partners/ access tools published by other partners.
Each institution in the example has a different internal setup, all of which are possible:
-
Partner A has a dedicated RCE server where the published tools are located, which is connected to the SSH gateway by an RCE connection. All other RCE users in the internal network are connected to this server.
-
Partner B has put all the tools directly on the SSH gateway instance.
-
In Partner C's network, some tools are located on the SSH gateway, but some tools are also published by users directly on their own machines. As long as they are connected to the SSH gateway also those tools can be published to the other partners.
-
Partner D has no tool server, instead the users’ computers connect directly to the relay server.
Note
Since RCE 10, the recommended connection type for secure connections are the SSH Uplink connections (cf. previous chapters). However, as the new Uplink connections do not yet support all features of the SSH Remote Access Connections (the publishing of workflows and the access to monitoring data is not possible by Uplink connections), the current release provides both types of connections. This chapter describe the usage of the SSH Remote Access Connections.
SSH connections provide a more secure alternative to the standard RCE connections and can be used to access tools remotely. The published tools are shown in the palette of the client's Workflow Editor (this may take a few seconds after connecting, as the tool list is fetched from the remote hosts every few seconds). From there, you can use them in your workflows just like your local components and tools. Differently from tools accessed by RCE network connections, in this case the component is shown to be executed on your local instance in the Workflow Execution wizard.
Also workflows that were published on the remote instance (for information about the publishing see section "Remote Tool and Workflow Access") are shown as components in the palette of the client's Workflow Editor in the group "SSH Remote Access Workflows" (if the client runs RCE 7.1 or newer). These remote workflows can be added to workflows and executed like local components/tools.
The RCE instance that publishes the tool, or another instance connected to it by RCE network connections, has to be configured as an RCE remote access server (see the "Configuration" section or the sample configuration file "Remote access server" for details).
Note
When configuring an SSH account using a key file, both server and client have
to run RCE 7.1 or newer. In RCE 10.0.0, only RSA Keys generated by the tool
puttygen using Windows-style line endings work. This is a known
issue with RCE 10.0.0 and will be fixed in an upcoming version of RCE.
When using Windows, the default settings of puttygen, which comes
bundled together with the popular SSH client putty, are sufficient.
When using Linux, you will have to install the tool puttygen.
Please refer to the documentation of your system for instructions on installing
that tool. After you have generated a key on Linux, you will have to convert it
to use Windows-style line endings. We recommend the tool todos for
this task. Both puttygen and todos are readily
available for most major distributions from the official package sources.
On the client side, SSH connections can be added in the "network" view by clicking "Add SSH Remote Access Connection". In the following dialog, enter the hostname and port of an RCE instance that provides an SSH server as well as the user name and the authentication data of an SSH account configured on this instance. Depending on the SSH account, you have to authenticate using a passphrase or by an RSA private key file. If your private key is protected by a passphrase, select the authentication type "Keyfile with passphrase protection", else select "Keyfile without passphrase protection".
The connections are shown in the "SSH Remote Access"->"SSH Remote Access Connections" subtree. They can also be edited, connected and disconnected in the "network" view. It is possible to store passphrases using the Eclipse Secure Storage Mechanism. However, the changes made here are not saved in the configuration yet, i.e. they will be lost when RCE is closed or restarted. To permanently add SSH connections, you can edit the configuration file (see the "Configuration" section for details).