Keep every application online, balanced and fast
Spread traffic across your servers at Layer 4 and Layer 7, terminate TLS, filter attacks with a WAF and fail over between appliances and sites. Play with the live demo below: it runs the same algorithms on simulated traffic.
- Requests/s
- 0
- Active connections
- 0
- Throughput (sim.)
- 0.0 MB/s
:443:443:3389Weights apply to weighted round robin. Moving cards show a sample of the requests; the counters include all of them.
Try the controls: switch the algorithm, drain or fail a server, or send a traffic spike.
From install to production
How it works
Install
Import the OVA or install the package on Debian 12 or Ubuntu 24.04.
Licence
Upload a signed licence file; it is verified offline.
Guided Setup
Pick an application profile; ports, checks and persistence are filled in.
Run
Watch health, drain servers for maintenance, automate with the API.
Features
Everything in one appliance
Layer 4 load balancing
Kernel-level balancing with IPVS in Direct Routing, NAT, tunnel (TUN) and SNAT modes, with firewall marks and nine scheduling methods including weighted round robin and least connections.
Layer 7 load balancing
HAProxy in HTTP or TCP mode with persistence, stick tables and seamless reloads. Every configuration is validated before it goes live.
SSL offload
Terminate TLS on the appliance with HAProxy or stunnel and keep every certificate in one store with expiry alerts.
Web application firewall
ModSecurity with the OWASP Core Rule Set in front of your web services, per virtual service.
GSLB
A built-in DNS responder that answers with a healthy site, using TCP and HTTP(S) monitors.
High availability
Clustered pairs with keepalived (unicast VRRP), floating IPs, configuration sync and Layer 4 connection-state sync.
REST API and lbctl
A versioned REST API with an OpenAPI description, scoped API keys, dry-run, batch and idempotency keys, plus the lbctl command-line tool.
Health checks and maintenance
Connect, negotiate, ping and script checks. Drain, halt or take a server offline without dropping users.
Rate limiting
Layer 7 protection rules that limit request and connection rates per client.
Backup, restore, signed updates
Encrypted and scheduled backups, checkpoints, roll back the last commit, and signed online or offline updates.
Real screens
Explore the console
Everything at a glance
- Virtual services, backend servers, connections and throughput.
- Drain, halt or take servers offline right from the list.
- Cluster role, licence and certificate expiry.

Services in one list
- Layer 7 HTTP, HTTP/2 and TCP modes, Layer 4 services alongside.
- Every commit is validated and can be rolled back.

Live traffic reports
- New connections per second, per service.
- Packets and data in and out.
- GSLB pools and members, stick tables.

See persistence at work
- Inspect which client is pinned to which server.
- Clear a table when you need to rebalance.

Backups you can trust
- Encrypted sensitive files, schedules, restore points.
- Deploy from a template to a new appliance.

Animated explainers
See the features in motion
Health-check failover
Servers that stop answering are taken out of the pool automatically and come back on their own when they recover.
- Requests are spread across three healthy servers.
- Server 2 stops answering its health check; requests sent to it time out.
- The load balancer removes it from the pool, so users only reach healthy servers.
- Health checks start passing again.
- Server 2 rejoins the pool automatically.
SSL offload
The load balancer terminates TLS with the certificate from its store, so your servers handle plain requests.
- Browsers connect over HTTPS; every request is encrypted (padlock).
- The load balancer holds the certificate and decrypts the request.
- Plain requests go to the servers, which no longer spend CPU on TLS.
WAF on the load balancer
ModSecurity with the OWASP Core Rule Set inspects requests before they reach the application.
- Normal requests are inspected and passed to the application.
- An attack is matched by the Core Rule Set and refused with an error page; the app never sees it.
GSLB across sites
The GSLB DNS responder only hands out sites whose monitors pass.
- Users get either site in the DNS answer and connect to it.
- Site A fails its TCP/HTTP monitors; GSLB stops handing it out.
- Site A passes its monitors again and is back in the answers.
High-availability pair
Two appliances share a floating IP; the secondary takes over if the primary fails.
- The primary owns the floating IP and handles traffic; the secondary stays in sync.
- The primary fails; the secondary claims the floating IP within seconds.
- Traffic continues through the secondary while the primary is repaired.
Under the hood
Architecture
Guided Setup
Ten application profiles
Pick your application and the profile fills in ports, mode, balancing method, persistence, health check and timeouts.
Microsoft Exchange 2016 / 2019
Client access over HTTPS plus an SMTP service, with an Exchange-aware health check.
Profile details →Remote Desktop Session Hosts
RDP on TCP 3389 with RDP-cookie persistence so users return to their session.
Profile details →IIS and general web servers
HTTPS web farms with source-IP persistence and an HTTPS health check.
Profile details →AD FS
Federation servers behind one name, checked through the AD FS probe endpoint.
Profile details →VMware Horizon (Unified Access Gateway)
Layer 4 NAT for the gateway ports over TCP and UDP with long source persistence.
Profile details →Citrix StoreFront
StoreFront servers with source persistence and a check of the store web page.
Profile details →Epic Hyperspace
HTTPS front end for Hyperspace web servers with one-hour persistence.
Profile details →S3-compatible storage (MinIO)
HTTP-mode balancing of object-storage nodes with the readiness endpoint as health check.
Profile details →SMTP relay
Ports 25 and 587 with STARTTLS kept end to end and a greeting-aware health check.
Profile details →LDAP and Active Directory
Direct Routing for LDAP and LDAPS so directory replies go straight back to clients.
Profile details →Hawkfend Load Balancer 13.1.2
Screenshot gallery







Platforms
Deployment options
VMware
OVA with four VMXNET3 ports and a console network wizard (2 vCPU, 4 GB RAM, 20 GB disk).
Bare-metal Linux
Tarball or .deb on Debian 12 or Ubuntu Server 24.04 LTS, amd64 or arm64.
Microsoft Azure
Bicep/ARM templates for a single appliance or an HA pair. Bring your own licence. Image and Marketplace soon
Hyper-V, KVM, Proxmox
Install the package in a Debian 12 or Ubuntu 24.04 VM. Ready-made images soon
Requirements
Minimum 2 cores, 4 GB RAM, 20 GB disk, 1 NIC. With the WAF, 8 GB RAM is recommended.
Models
Editions
Every model includes every feature on this page. Choose by form factor and capacity.
| Model | Form factor | Intended for | Throughput | Availability |
|---|---|---|---|---|
| HLB-100 | Hardware appliance | Small deployments | To be published | Contact sales |
| HLB-300 | Hardware appliance | Mid-size deployments | To be published | Contact sales |
| HLB-500 | Hardware appliance | Large deployments | To be published | Contact sales |
| HLB-V | Virtual / software | Any VM, cloud or Linux server | To be published | Software, available now |
Placeholder Capacity figures are placeholders. Compare models →
Coming soon
- Automatic ACME certificate issuance and renewal on Linux appliances (client built, final step in progress)
- Single sign-on, two-factor sign-in and custom roles for the appliance console
- Automatic WAF rule-set updates
- OCSP stapling and selective restore
- Prometheus-style metrics and an SNMP agent
- Ansible and Terraform integrations
- Azure image, Marketplace listing and pay-as-you-go billing
- Ready-made images for Hyper-V, KVM and Proxmox; RHEL support
Load balance your first service today
Request a trial licence, install on a Linux server or virtual machine you already have, and decide with your own traffic.