Workload management through SYMREST

REST-based workload is available only with IBM® Spectrum Symphony Advanced Edition, using the REST-based SYMREST server. Submit and manage workload from IBM Spectrum Symphony clients to your cluster through the REST-based SYMREST service, or submit and manage IBM Spectrum Symphony applications and workload (sessions and tasks), through these RESTful API methods.

For client submissions, SYMREST is a long-running system service, which runs on management hosts in the cluster and interacts with your RESTful API clients for client submissions. It serves as a REST proxy server between the RESTful API client and the cluster to create client sessions and submit workload to an application in the cluster. With SYMREST, clients can submit and receive workload without requiring changes to an existing application service.

For IBM Spectrum Symphony applications and workload, the SYMREST is a long-running system service, which runs on management hosts in the cluster and allows you to monitor and manage commands for your applications, application profiles, and workload (which includes IBM Spectrum Symphony sessions and tasks). The RESTful APIs commands interact with SYMREST server.

With REST-based workload submission for clients, you can derive the following benefits:
  • Secure communication with TLS to protect the integrity and confidentiality of data transmission. TLS communication for RESTful APIs follows your security setting for web servers, which is enabled during installation by default. With this setting, the RESTful API client uses a built-in self-signed certificate; this self-signed certificate is meant only for testing purposes. For your production environment, you must use a properly chained certificate that is issued or signed by a trusted certificate authority. See Securing communication with RESTful API workload for SYMREST.
  • For client workload submission, control access to your cluster by defining only some applications for the workload. This configuration restricts application access and is useful to limit cluster access, especially if your clients are located in the cloud. See Managing workload from a RESTful API client.
  • Increase scalability by adding more instances of the SYMREST service. With multiple service instances, you can take advantage of options to balance workload requests among a list of available SYMREST servers, as follows:
    • (Linux® only) Use ETCD servers to query ETCD to determine the least-loaded SYMREST server. To balance the load with ETCD servers, an ETCD installation is a prerequisite; take steps to install and configure ETCD in your cluster. See Balancing workload with ETCD.
    • Use random selection, wherein a client connects to a SYMREST server in the cluster at random. See Balancing client workload with random selection.
  • (Linux only for client workload submissions) Enable single sign-on authentication through OpenID for client connections. OpenID is an open standard that enables users to authenticate themselves to multiple entities without managing multiple accounts or sets of credentials. IBM Spectrum Symphony supports OpenID Connect 1.0, which is an authentication layer on top of the OAuth 2.0 authorization framework. To use OpenID authentication for your clients, an OpenID identity provider (IdP) is a prerequisite; set up your IdP according to your IdP's instructions and complete the OpenID client registration process. See Configuring OpenID authentication for RESTful API client workload.
Restriction: REST-based workload submission supports session creation, task creation, task result retrieval, and session closing. Common data update, symexec execution, and AdminConnection APIs are not supported.
Tip: IBM Spectrum Symphony Developer Edition provides a sample Java™ client, CloudProxyClient, which you can use to test workload submission from a cloud client. The CloudProxyClient sample provides one client and is available at:
  • %SOAM_HOME%\version\samples\Java\CloudProxyClient on Windows.
  • $SOAM_HOME/version/samples/Java/CloudProxyClient on Linux.

For instructions on using the CloudProxyClient sample, refer to the accompanying readme file.