IBM zEDC to increase max length of output buffer to more than 1 MB

In IBM cloud tape connector(CTC) product , we use the zEDC callable service FPZ4ABC to compress data. This service supports a maximum input and output buffer size of 1 MB. For performance reasons, CTC always invokes this service with the maximum input buffer size, assuming the compressed output will fit within the 1 MB output buffer. However, in some customer cases, when the data is not compressible, the output of a 1 MB input can exceed the maximum output buffer size supported by FPZ4ABC, resulting in RC=12, reason code 206.

We had previously raised case TS009389983 regarding this issue, and the zEDC team recommended implementing a retry mechanism. We have implemented this in the CTC code by retrying the compression with an input buffer smaller than 1 MB when the data is not compressible.

Recently, however, we encountered a highly business‑critical customer case where RC=12/206 occurred at a very high frequency. Because the failures happen at scale, each retry leads to significant performance degradation, especially given that our customers process data volumes in terabytes.

This appears to be a boundary-condition issue acknowledged by the zEDC team—that incompressible data may expand beyond its original size. Given this, there is a functional gap: if the service supports a 1 MB input buffer, the output buffer should ideally allow for a larger size.

We request increasing the maximum output buffer size from 1 MB to 2 MB, so that in cases where the data expands rather than compresses, the output can still be accommodated without causing RC=12/206 failures.