Looking at Netskope's ZTNA for a 200-user Google Workspace shop. Main requirement is strict, integrated DLP for Drive/Gmail.
Ran a basic comparison of key features against the pure-play ZTNA vendors. Netskope's strength is the unified Client/Clientless access combined with its existing CASB DLP engine.
**Key points:**
* **ZTNA + DLP Integration:** DLP policies from the CASB side can apply directly to ZTNA tunnels. No need for separate policy engines.
* **Client Deployment:** Requires the Netskope client. No agentless option for managed devices.
* **Google Workspace Specifics:** DLP for Google Drive is robust. Real-time coaching for users works in browser.
* **Pricing Model:** Per-user, includes ZTNA, SWG, CASB, DLP. Can get expensive if you only need ZTNA+DLP.
**Example of a simple DLP rule that would apply to both ZTNA and direct web traffic:**
```
rule "Prevent Google Drive PII Upload via ZTNA" {
actions = [block, alert]
conditions {
service = netskope_ztna // Or 'google_drive' for direct access
content.category = "PII"
file.type = [".doc", ".pdf", ".csv"]
}
}
```
Question for others: In a real deployment, how does the ZTNA session handling perform when the DLP deep inspection is enabled on large file transfers? Any latency spikes?
- bench_beast
Benchmarks don't lie.
I'm a FinOps lead at a 250-person financial services firm, and we've been running Zscaler Private Access for our Google Workspace environment for about two years, specifically for its integrated DLP controls on Drive and Gmail.
* **Google Workspace DLP Integration Depth:** Zscaler's DLP engine uses the same exact policy set for ZTNA and direct web traffic via its Client Connector. The key detail is you can apply DLP scanning to the TLS decryption for ZTNA tunnels targeting `drive.google.com` and `mail.google.com` specifically. In our implementation, this catches PII before it hits Drive via a native app syncing through the tunnel. The policy console is unified, but initial rule translation from our legacy tool took about 40 hours of tuning.
* **Real-World ZTNA Session Performance with DLP:** With DLP inspection enabled on the ZTNA path for Google services, we measured a consistent 8-12% increase in latency (adding 25-40ms) for document operations in Drive. Throughput for large file uploads via the tunnel decreased by roughly 15% compared to a DLP-off policy. For our 200 users, we've never saturated it, but the system clearly prioritizes security over speed when full-content inspection is triggered.
* **Total Cost of Ownership for 200 Users:** The list price for Zscaler Private Access (ZPA) with the Advanced DLP add-on was approximately $9.50 per user per month on a one-year term. The hidden cost is the compute required for the App Connectors (their gateway proxies); we run three `c5.xlarge` instances in AWS us-east-2, which adds about $290/month. The all-in cost for us is closer to $11/user/month, not the often-quoted $7-9.
* **Deployment and Management Friction:** The Client Connector deployment via Google Workspace was straightforward. The significant effort was in configuring the App Connectors for high availability and defining granular service segments for each Google API endpoint. You'll spend a week defining which micro-services (like `docs.google.com` vs `drive.google.com`) get which access policies. Their support was responsive during setup (sub-30 minute replies) but now operates on a 4-8 hour SLA for standard tickets.
I would recommend Zscaler Private Access if your primary requirement is a deeply integrated, single-console DLP policy engine for ZTNA and web traffic to Google Workspace. If your budget is under $8/user/month all-in and you can tolerate a separate DLP console for non-ZTNA traffic, tell us your exact compliance framework and whether you need to support unmanaged devices, as that changes the landscape.
Every dollar counts.
Great breakdown on the policy unification, that sample rule is exactly the kind of thing that makes the Netskope pitch so tempting. I'm curious though about your experience with the client side of things. You mentioned it requires the Netskope client and no agentless option for managed devices. We're a similar size shop, about 180 users, and we tried a Netskope POC last year. The DLP scanning on Drive was solid, but we ran into nasty latency on the ZTNA tunnel when the client was doing TLS decryption for large file syncs through the native Drive app. The real-time coaching pop-ups in the browser were fine, but background syncs would just silently fail or time out.
Did you see anything like that in your testing? Or did you focus mostly on browser-based access? I'm trying to figure out if that's a config thing or a fundamental tradeoff with the per-user pricing model. It gets expensive fast if you have to double up on tools just to handle the native app workflows.
cost first, then scale
That unified policy example is interesting. I've been reading about the "clientless" access option mentioned in Netskope's marketing. In your testing, does that clientless mode work for Google Workspace at all? If it does, would it bypass the DLP scanning that happens through the full client's tunnel? I'm trying to understand if the integration is truly seamless or if there are separate policy paths.
That's a critical distinction, and your instinct is correct. The clientless mode does work for browser-based access to Google Workspace, but it operates on a completely separate policy path.
Clientless access uses a reverse-proxy architecture for unmanaged devices, which means the DLP engine scanning that traffic is the **inline CASB** component, not the ZTNA engine. While Netskope's policy console is unified, the underlying enforcement points differ. A DLP rule defined for ZTNA client traffic will not automatically apply to clientless sessions unless you explicitly bind it to the CASB service as well.
The practical risk is a compliance gap: a user on an unmanaged device accessing Drive via clientless mode could bypass the specific ZTNA DLP policy you've tuned for native app syncs. You must duplicate and maintain the policy logic across both service definitions, which adds administrative overhead and potential for drift.
—LJ