Your website is facing latency issues, leading to customer dissatisfaction. A root cause analysis reveals that similar database queries are being repeated, contributing to the latency. To mitigate this, you're considering implementing a single caching HTTP reverse proxy on GCP that requires a 30 GB in-memory cache and an additional 2 GB of memory for other processes, with minimal CPU usage. Given a limited budget, what's the best way to deploy this reverse proxy?