Hey folks! Been diving deep into our cloud spend reports and realized our Zscaler Internet Access (ZIA) bandwidth costs were creeping up. We migrated from a traditional NGFW setup a while back, and I was curious: are we *actually* saving money, or just moving cost centers?
So, I built a quick Python script to pull data from our ZPA and old firewall logs (exported as CSVs) and compare the bandwidth-based billing. The goal was to get a clear, monthly cost projection side-by-side.
Here's the core of the script – it's pretty simple but super effective:
```python
import pandas as pd
import matplotlib.pyplot as plt
def calculate_monthly_cost(bandwidth_gb, price_per_gb):
"""Calculates monthly cost based on GB usage."""
return bandwidth_gb * price_per_gb
# Load data (simplified example)
df_zia = pd.read_csv('zia_bandwidth_log.csv')
df_old_fw = pd.read_csv('legacy_fw_bandwidth.csv')
# Aggregate monthly bandwidth in GB
zia_monthly_gb = df_zia.resample('M', on='timestamp').sum()['gb_sent']
old_fw_monthly_gb = df_old_fw.resample('M', on='timestamp').sum()['gb_total']
# Apply pricing (using our negotiated rates)
zia_rate = 0.08 # $ per GB
old_fw_rate = 0.22 # $ per GB (includes estimated support & hardware amortization)
zia_cost = zia_monthly_gb.apply(lambda gb: calculate_monthly_cost(gb, zia_rate))
old_fw_cost = old_fw_monthly_gb.apply(lambda gb: calculate_monthly_cost(gb, old_fw_rate))
```
**What I found:**
* Our ZIA costs are about **35% lower** on raw bandwidth billing.
* The big hidden savings came from **reduced overhead**: no more HA pair licensing, fewer admin hours on rule updates, and way less time troubleshooting VPNs.
* The script also highlighted some "chatty" dev instances we didn't know about – bonus!
**Key takeaway:** The pure per-GB cost is better, but the real win is in operational efficiency. The script is now part of our monthly FinOps check-in.
Has anyone else done a similar comparison? I'd love to hear what metrics you used or if you factored in things like threat prevention tiers. The script is a work in progress, happy to share the full version if anyone's interested.
Keep deploying!
Keep deploying!