Enterprise networks today are a patchwork of proprietary hardware appliances — separate boxes for routing, firewalling, VPN termination, load balancing, and traffic inspection. Each with its own management plane, licensing model, and upgrade cycle. Each from a different vendor. Each a potential point of failure.
The result? Operational complexity that scales faster than your business. Every new service requires a new appliance, a new integration, a new maintenance window. Your team spends more time managing the network than innovating on it. And when a bottleneck appears at 100 Gbps, the only answer is a forklift upgrade to the next generation of proprietary silicon.
This model is broken. The industry knows it. SDN was supposed to fix it — but most SDN solutions sacrifice performance for flexibility, dropping to a fraction of line speed when real traffic processing begins. You shouldn't have to choose between programmability and performance.
The OM-GS-1000 consolidates your entire network infrastructure into a single software-defined platform. Router, firewall, VPN gateway, load balancer, traffic inspector — all running concurrently, all at line speed, all on standard hardware.
Replace multiple proprietary appliances with one unified stack. Route, filter, encrypt, balance, and inspect — all in a single deployment.
Kernel-bypass architecture (DPDK) delivers 200 Gbps L3 forwarding per core with up to 80% headroom remaining. No compromises between features and throughput — every function runs at wire speed.
Performance scales linearly with CPU cores. Need more throughput? Add cores, not appliances. Intel benchmark-proven from 1 to 6+ cores with predictable, linear growth.
Complete network stack from Ethernet switching through application-layer inspection. Pre-integrated applications for every layer — no gaps, no bolt-ons, no integration tax.
Hardware-accelerated cryptographic operations via Intel QuickAssist Technology. 4 cores (2 SW + 2 QAT) deliver 96 Gbps — surpassing 6 software-only cores at 92 Gbps while freeing CPU capacity.
Built-in microsegmentation, encrypted east-west traffic, and identity-aware access policies. Every packet is inspected, every flow is authorized, every session is encrypted.
The GateSpeed architecture bypasses the kernel entirely, moving packet processing into user space via DPDK. Each layer of the stack is purpose-built for performance.
Based on controlled testing at Intel Labs using the GateSpeed software stack on Intel Xeon processors. Values are approximate readings from the published white paper charts.
AES-GCM encrypted tunnel performance measured across packet sizes from 64 bytes to 1455 bytes, scaling from 1 to 6 CPU cores. Values below are approximate readings from the Intel Labs white paper charts. Results demonstrate near-linear scaling.
| Packet Size | 1 Core | 2 Cores | 3 Cores | 6 Cores |
|---|---|---|---|---|
| 64 B | 8 Gbps | 16 Gbps | 25 Gbps | 48 Gbps |
| 128 B | 15 Gbps | 30 Gbps | 45 Gbps | 70 Gbps |
| 256 B | 22 Gbps | 42 Gbps | 60 Gbps | 85 Gbps |
| 512 B | 30 Gbps | 55 Gbps | 75 Gbps | 92 Gbps |
| 1024 B | 38 Gbps | 65 Gbps | 85 Gbps | 97 Gbps |
| 1455 B | 42 Gbps | 70 Gbps | 90 Gbps | 100 Gbps |
HTTP and HTTPS reverse proxy throughput tested with varying concurrent connections and file sizes on the GateSpeed platform. 2 MB results verified from the GateSpeed Intel webinar; 10 KB values are approximate readings from the white paper.
| Connections | File Size | Throughput | CPU Usage |
|---|---|---|---|
| 5,000 | 2 MB | 85 Gbps | 2 cores @ 95% |
| 10,000 | 2 MB | 82 Gbps | 2 cores @ 96% |
| 20,000 | 2 MB | 82 Gbps | 2 cores @ 95% |
| 40,000 | 2 MB | 82 Gbps | 2 cores @ 95% |
| 5,000 | 10 KB | 90 Gbps | 4 cores @ 95% |
| 10,000 | 10 KB | 88 Gbps | 4 cores @ 95% |
| 20,000 | 10 KB | 88 Gbps | 4 cores @ 95% |
| 40,000 | 10 KB | 82 Gbps | 4 cores @ 95% |
| Connections | File Size | Throughput | CPU Usage |
|---|---|---|---|
| 5,000 | 2 MB | 74 Gbps | 4 cores @ 95% |
| 10,000 | 2 MB | 68 Gbps | 4 cores @ 95% |
| 20,000 | 2 MB | 66 Gbps | 4 cores @ 95% |
| 40,000 | 2 MB | 64 Gbps | 4 cores @ 95% |
| 5,000 | 10 KB | 71 Gbps | 4 cores @ 95% |
| 10,000 | 10 KB | 67 Gbps | 4 cores @ 95% |
| 20,000 | 10 KB | 63 Gbps | 4 cores @ 95% |
| 40,000 | 10 KB | 61 Gbps | 4 cores @ 95% |
All benchmarks performed at Intel Labs, Spring 2024. Results based on the GateSpeed software stack running on Intel Xeon processors with Intel E810 100GbE NICs. Actual performance may vary based on configuration and workload. See Intel white paper for full methodology.
Download Intel White Paper →The OM-GS-1000 is fully software-defined — every network function is deployed, configured, and managed through APIs and automation. No truck rolls, no hardware swaps, no maintenance windows.
Every configuration change — from firewall rules to VPN tunnels to load balancer pools — is driven through a RESTful API. Integrate with your existing CI/CD pipelines and Infrastructure-as-Code tools.
Add or remove network functions at runtime without rebooting or disrupting traffic. Need a new VPN endpoint? Deploy it in seconds. Decommissioning a service? Remove it without downtime.
Run multiple isolated network environments on a single platform. Each tenant gets dedicated resources, policies, and monitoring — perfect for managed service providers and large enterprises.
Security isn't bolted on — it's built into every layer of the stack. From wire-speed encryption to deep packet inspection, the OM-GS-1000 makes zero-trust networking practical at scale.
Stateful inspection, application-aware filtering, and DDoS mitigation — all at line speed. Identity-based policies, microsegmentation, and behavioral analysis at the edge.
Distribute traffic across backends with health checking, session persistence, and weighted routing. Sub-millisecond decisions across TCP, UDP, HTTP, and gRPC.
Per-flow metrics, latency histograms, and anomaly detection — exported to Prometheus, Grafana, or your SIEM in real time.
High-bandwidth, low-latency fabric for GPU clusters. RDMA-capable, lossless Ethernet for distributed training workloads that demand predictable network performance.
See AI Platform →Deploy at the edge with full network stack capability in a compact footprint. Content caching, SSL offload, and traffic optimization close to end users.
User Plane Function (UPF) acceleration, network slicing, and multi-access edge computing. Carrier-grade reliability with software-defined agility.
Consistent networking across on-prem, colo, and public cloud. Overlay networking, site-to-site VPN, and unified policy management regardless of where workloads run.
Whether you're replacing a single appliance or re-architecting your entire network, the OM-GS-1000 meets you where you are.
Start by replacing a single appliance — your VPN gateway, load balancer, or firewall. Same functionality, better performance, lower cost. Zero disruption to existing architecture.
Collapse multiple network functions onto a single OM-GS-1000 instance. Reduce rack space, power consumption, and management overhead while gaining unified telemetry and policy control.
Re-architect with software-defined networking end-to-end. API-driven automation, multi-tenant isolation, and centralized management across all sites and environments.
From drop-in appliance replacement to full SDN migration — the OM-GS-1000 delivers line-speed networking on standard hardware. Let us show you what it can do.
