COMPARISON OF THE 1C LOAD TEST TECHNOLOGY AND MULTI-AGENT SIMULATION BASICS

The 1C platform-based IT systems load tests technology is described. Principles of the 1C methodology create load generation robots using the builtin 1C language. Other distinctive features of 1C methodology are realistic, not stressful scenario and qualitative passing of multistep business process, not chaotic load generation. The scales of usual 1C load tests up to 400…700 users are given. The examples of large tests up to 25 000 users in the same base are described. The main task of load tests is to reduce technological risks. The other tasks are: making marketing promotion, making hardware sizing and to looking beyond the horizon. Definitions of the multi-agent approach are mapping to the 1C methodology practical implementation and to the technology toolkit. Each load generating client is an independent agent of multi-agent simulation modeling. There is difference between agents and actors. Agents simulate business functions but actors are a part of technological toolkit and do service functions only. Developers in testing teams create and debug exactly agents using templates. Actors are provided by toolkit vendor, first of all – the 1C Company. It is shown, how this technology covers under the classification of multi-agent simulation modeling approaches. Relations between the actual technique and the theoretical definitions demonstrate that 1C load tests methodology is a special method of multi-agent simulation modeling indeed. Classifiers’ applicability additionally confirms this concept. It is shown, that 1C platform-based load tests technology is an either systems with simple reflex agents or systems with model-based reflex agents. Agents are active objects. Agents are real software. Finally, these systems are homogeneous or heterogeneous, with continuous time and model space represents business process maps.