1. Documentation
2. 2024.1 (Latest) 2024.0 10.3 10.2 10.1 10.0 9.10 9.9 9.8 9.7 9.6 9.5 9.4 9.3 9.2 9.1 9.0 8.2 8.1 8.0 7.6
3. Hub
4. Endpoints
5. Real-Time Scoring

Real-Time Scoring

Terminology change

The documentation in this section has been mostly dedictated to Web API endpoints, with what used to be called Real-Time Scoring described as an afterthought.

But Real-Time Scoring still exists, although the terminology used now is somewhat different. In this section, whenever we talk about a scoring agent or a regular endpoint, we are in fact talking about Real-Time Scoring. The language is different, but the technology is the same.

In contrast, the new and improved technology is called Web API endpoints. As described in the introduction, Web API agents can communicate with Altair AI Hub, and this communication sets them apart from regular scoring agents and the whole concept of Real-Time Scoring.

From a developer perspective, the main difference between the old and new is ease of use: Web API agents can automatically update new versions of your processes.

From a documentation perspective, it makes sense to emphasize the easy-to-use technology that is likely to see greater use -- hence the emphasis on Web API endpoints. However, this document aspires to bring greater clarity to Real-Time Scoring agents.

Edge computing

Despite the above comments, the Real-Time Scoring technology is not obsolete! If, as in the example of edge computing, you need to obtain results from a model without having access to Altair AI Hub, you need a regular scoring agent / Real-Time Scoring.

As discussed in Create an endpoint, use of an external scoring agent independent of Altair AI Hub requires more effort, both because you have to install the endpoint infrastructure (install a scoring agent) and because the endpoint processes are not updated automatically. Any time you want to update the endpoint and request new results, you have to take the following steps:

1. download a deployment ZIP from Altair AI Hub,
2. upload the ZIP file to your external endpoint infrastructure,
3. place the ZIP in the deployments folder,
4. restart the scoring agent, and
5. request new results.

However, the benefit is also clear. You can have your endpoint anywhere you like: on your laptop, on a remote server, or on your Raspberry Pi! You no longer depend on Altair AI Hub.

Example

Let's say, for example, that you want to create an endpoint / Real-Time Scoring deployment on your laptop for use in an air-gapped environment (no internet!) You will need the internet to install the endpoint infrastructure and create an endpoint, but once that's done, you're going to be disconnected.

You're looking for a simple installation method not involving Docker or Kubernetes, so you choose the Standalone method.

Once you have installed the scoring agent, the five steps described above take the following form:

1. download a deployment ZIP from Altair AI Hub, by one of the following methods:

   • Create a deployment ZIP within the Altair AI Hub user interface, or
   • Use a REST API to create a deployment ZIP

2. upload the ZIP file to your external endpoint infrastructure,

   • not necessary, because the scoring agent is already on your laptop

3. place the ZIP in the deployments folder,

   • as described in Install a standalone scoring agent, point (5)

4. restart the scoring agent, and

   • as described in Install a standalone scoring agent, point (6)

5. request new results.

   • as described in Install a standalone scoring agent, point (7)