There is no definite limit "on my own" (at least I do not know). This mainly depends on how the data is stored in memory, what kind of server load and how many resources are available. For the most part, contiguos of memory blocks (RAM) are likely to become the scarce resource, so we should focus on that.
βHow data is stored in memoryβ refers to the fact that for greater throughput, some transports (buffered, framed) tend to allocate more memory and larger blocks than others. Depending on the implementation of the language, this process can be implemented more or less efficiently in terms of the cost of memory.
If you really plan to transfer large blocks of data, you should also look at other parameters, for example
- blocking data
- sending / returning only the URL resource or local network through the service, and not all data
source share