By what other methods, besides multithreading, can the efficiency of a report-generating service that performs 4 parallel HTTP requests to different systems be increased?

In addition to multithreading, to improve the efficiency of a service performing 4 parallel HTTP requests, asynchronous programming and non-blocking I/O can be used. In Java, this is implemented using CompletableFuture, reactive libraries (such as Reactor or RxJava), or NIO.

The asynchronous approach allows not to block the thread while waiting for a response from external systems, which increases overall performance and scalability.

Example with CompletableFuture:

CompletableFuture<String> request1 = CompletableFuture.supplyAsync(() -> makeHttpRequest(url1));
CompletableFuture<String> request2 = CompletableFuture.supplyAsync(() -> makeHttpRequest(url2));
CompletableFuture<String> request3 = CompletableFuture.supplyAsync(() -> makeHttpRequest(url3));
CompletableFuture<String> request4 = CompletableFuture.supplyAsync(() -> makeHttpRequest(url4));

CompletableFuture.allOf(request1, request2, request3, request4).join();

String result1 = request1.get();
// and so on

You can also consider caching the results of requests if the data does not change often, to reduce the number of calls to external systems.

Another option is to use a connection pool and optimize HTTP client settings to reduce delays.