SKUDONET Enterprise Edition 10.4.0 is now available. Its main addition is SNAT: you can now choose the source IP address the load balancer uses when it forwards traffic to each backend.

The release also extends Last Hop to TCP SSL and L4xNAT farms. It adds a resource manager for management and protection services, and brings a few usability improvements and several fixes in farms and Last Hop.

What’s new in SKUDONET Enterprise Edition 10.4.0?

  • SNAT for backends, with a selectable source IP per backend
  • Last Hop support in TCP SSL and L4xNAT farms
  • A new resource manager: skd-resource-governor
  • Search in selection lists and by columns in tables
  • Better handling of network interfaces whose link comes up late
  • Six fixes affecting L4xNAT, TCP SSL, HTTP farms and Last Hop

SNAT: choose the source IP for each backend

With SNAT, administrators select the source IP address used when forwarding traffic to each backend. That gives you more control over how the load balancer connects to your servers, and it simplifies firewall allowlisting, since the traffic arrives from an address you picked.

It also helps direct return traffic through the load balancer. And because every backend can have its own source IP, you can adapt the setup to what each network requires.

We explain how it works and how to configure it, step by step, in the technical article Source NAT (SNAT) in SKUDONET Enterprise Edition 10.4.0: Set the Source IP per Backend.

Last Hop now also covers TCP SSL and L4xNAT farms

Last Hop arrived in Enterprise Edition 10.2.2 for HTTP and HTTP/2 farms. With 10.4.0, it also works in TCP SSL and L4xNAT farms, so control over the return path reaches more application protocols.

The load balancer learns where incoming connections arrive from and uses that to send the response through the corresponding network interface and the previous network device. It’s especially useful in environments with several gateways or firewalls.

If you want the background on asymmetric routing and how this works, it’s all in our Last Hop technical article.

A resource manager for management and protection services

10.4.0 introduces skd-resource-governor, a new system resource manager. It organizes the system processes and controls how much CPU, memory and how many processes the management and protection services can consume.

Those services are now separated from the traffic processing runtime, which helps contain resource spikes and reduces their impact on application delivery. The manager also shows how resources are allocated, which is useful for troubleshooting and capacity planning.

Improvements

Search in selection lists and tables. Selection lists now have a search option, and tables can be searched by column to narrow results using specific attributes. In larger deployments, finding the object you need takes less time.

Interfaces that come up late. We’ve improved how SKUDONET handles an interface when its physical link becomes available (netplug), including when a switch port comes up after the system has already started. The interface configuration is applied correctly, and connectivity recovers better after a delayed link.

Bug fixes

L4xNAT farms

  • Backends in down or maintenance state are now handled consistently. You can take servers out of service for maintenance, and unavailable backends are no longer treated as valid destinations for new connections.
  • FarmGuardian health checks now evaluate backends in down state correctly, so recovered servers are detected and their status is updated.
  • Port validation for SCTP configurations has been corrected, and valid SCTP settings are now accepted.

TCP SSL farms

  • FarmGuardian integration now handles health check results correctly. The availability detected by monitoring matches the backend status the farm uses to distribute traffic.

HTTP farms

  • L7 Smart Routing now applies routing rules correctly in configurations that use only HTTP farms.

Last Hop

  • A stopped farm that overlaps with another farm’s configuration no longer interferes with Last Hop in the farms that are active.

If you want to see SNAT and the rest of the changes on your own infrastructure, download the evaluation: