Skip to content

On…

OpenPulse

Private access

Your servers answer only to the people your policy names.

Private access gives your organization its own encrypted network. Devices and servers join through OpenPulse sign-in, everything is denied until a rule allows it, and sensitive access opens for a set time after a second person approves.

By request · one network per organization

One request, from the ask to the expired certificate.

This is the whole path of one request at Sorvenna Logistik GmbH, a made-up freight company. Nothing on it is standing: the grant, the open path and the SSH certificate end together.

  1. 09:02

    Lena asks

    An engineer writes: “I need the database for 30 minutes. Last night’s invoice run failed.” The request names db-prod, the reason and the duration.

  2. 09:04

    Jonas approves

    The policy names him as an approver for db-prod. He confirms with his passkey, so a stolen session could not have approved.

  3. 09:04

    The path opens

    Lena’s laptop reaches db-prod, and her SSH session gets a certificate for the account the policy allows.

  4. 09:34

    Everything ends

    The grant runs out. The path closes and the certificate expires with it.

  5. After

    The record stays

    The request, its reason, the approval and the certificate are entries in Sorvenna’s access record.

An encrypted network of your own.

Each organization gets its own private network. Traffic between a device and a server is encrypted with WireGuard from end to end, and devices meet through a relay, so a server needs no port open to the internet.

One network per organization
Your devices and resources share a network with nobody else’s.
Closed by default
A new network allows nothing. A path exists only because a rule in your policy names it.
Only what you route
The tunnel carries traffic for your private resources. Everything else on the device goes its normal way.

People join by signing in. Servers join by name.

A person brings a device onto the network by signing in with OpenPulse, and only an active member of the organization is admitted. A server or internal app joins as a named resource.

Bound devices
A device proves a key it holds, so the network knows the device as well as the person.
Named resources
A server joins with a setup key that works once and expires within the hour.
Network devices in the Console
Every device with its owner, system and last sign-in, and a button to remove it.

One document says who reaches what.

The policy is a single document per organization, written as YAML or JSON. A rule goes from groups, people or bound devices to resources, with ports and protocol. You can keep it in your own repository.

See the difference first
Before a policy applies, the Console shows what it changes.
History
Every applied policy is kept, with who applied it and when.
Groups from your directory
Rules name the same groups your identity provider manages.

Sensitive access needs a reason, a second person and an end.

For a sensitive resource nobody holds standing access. A person asks with a reason and a duration, an approver named in the policy confirms with a fresh passkey step, and the path opens until the time is up.

  1. 01

    Ask

    Pick the resource, give a reason and a duration.

  2. 02

    Approve

    A second person confirms with their own passkey.

  3. 03

    Reach

    The path opens for the set time, and an admin can end it at once.

  4. 04

    End

    The grant closes on its own. Open, approved and ended requests stay listed in the Console.

No long-lived keys sit on your servers.

A granted SSH session gets a certificate signed by your organization’s certificate authority. It names the accounts the policy allows and ends with the grant, four hours at most. Your servers trust the authority, so no personal key is ever copied onto them.

Every step is on the record, and leaving removes everything.

Every join, policy change, request, approval and certificate is written to an access record per organization, each entry chained to the last by hash. When a person leaves the organization, their devices, grants and certificates leave the network with them.

What the person who runs the servers asks next.

Can OpenPulse read what travels through the tunnel?
No. WireGuard encrypts from the device to the server, so what passes a relay is already encrypted. OpenPulse adds no cryptography of its own to the tunnel.
How is an SSH certificate taken back?
It is not taken back, it runs out. A certificate lasts as long as its grant, four hours at most, and there is no separate revocation list. That is why certificates are short-lived.
Which parts are yours, and which are open source you build on?
The tunnel is WireGuard, the network builds on NetBird’s open source data plane, and SSH certificates come from step-ca, an open source certificate authority. OpenPulse adds the sign-in, the policy, the approvals, the certificates’ link to a grant and the record.

Next: planned, and not part of the product yet.

  • Device posture in policy

    A rule that admits a device only when its operating system and client are current. Today the Console shows these facts for each device; the policy does not act on them yet.

  • Protective DNS

    Name lookups on the private network checked against blocklists of known malicious domains, with no log of what people look up.

Where it stands today.

  • The network runs on WireGuard and builds on NetBird’s open source data plane.
  • Our use of WireGuard, NetBird and step-ca has not been independently audited.
  • Private access opens by request, one organization at a time.

How security and privacy work

Close your servers to everyone but the right person.

Tell us what you need to reach and who needs to reach it. We set up your network and your first policy with you.