If you have 2^143 * 200 = 2230074519853062314153571827264836150598041600 PS3 and 6 months. And the electricity to power them, let's say about 100W per PS3.

You could look up the algorithms here: https://www.math.auckland.ac.nz/~sgal018/crypto-book/crypto-book.html
BSGS is in chapter 13.3
When memory is limited - Distributed Kangaroo in chapter 14.6

Pollard Rho can be modified for an interval shorter than the whole group, at the cost of more group operations, i.e. for range 1 to 2^160 a possible solution would take 10*3*2^80 group ops plus 10*2^16 group points additional memory.

There are other probabilistic algorithms when private key is in a range significantly smaller than the size of the whole group - Pollard Kangaroo, and Gaudry-Schost algorithms.

Pollard Kangaroo with four kangaroos (the optimal) and distinguished points (DP) has expected work 1.715*2^80.

Four set Gaudry-Schost algorithm with DP gives 1.661*2^80 group operations, and is easy to parallelise.

There might be additional memory requirements for both methods, negligible compared to the enormous cost of BSGS.

This is correct only for BSGS (Baby-Step-Giant-Step).

Using Pollard Rho method, the expected work is 3*2^80 group operations with almost zero memory requirements.

Note that unlike BSGS  this method is probabilistic, and might fail with very low probability (on the order of 2^-160).

One can improve the algorithm using Distinguished Points, bringing the expected work down to 1.253*2^80 group operations, using both less memory and less group operations (on average) than BSGS.