.\" Ouroboros man pages CC-BY 2017 - 2021
.\" Dimitri Staessens <dimitri@ouroboros.rocks>
.\" Sander Vrijders <sander@ouroboros.rocks>

.TH OUROBOROS 8 2020-11-29 Ouroboros "Ouroboros User Manual"

.SH NAME

Ouroboros - the free IPC subsystem

.SH SYNOPSIS

\fIirmd\fR [--stdout] [--version]

\fIirm\fR [ipcp] <command> <args>

.SH DESCRIPTION
Ouroboros is an Inter-Process Communication (IPC) subsystem consisting
of an IPC Resource Manager (IRM), a number of IPC Processes (IPCPs)
and a library. It provides an API to establish communication between
processes, regardless of whether they are running locally or remote.

See \fBouroboros-tutorial\fR(7) to get started.

This page provides an overview of the command options for the
\fBirmd\fR and associated \fBirm\fR toolset. You can learn more about
individual commands with "irm [command] help".

.SH IRM DAEMON
The \fBIPC Resource Manager (IRM)\fR is the core component of the
Ouroboros subsystem. It manages \fIIPC processes\fR and the
datastructures used for communication over Ouroboros. Before using
Ouroboros, an IRM daemon must be running on your system. The IRM
daemon can be started as a process at the command line (as a
superuser):

$ irmd

or, if your system uses systemd, Ouroboros can be started as a service:

$ systemctl start ouroboros

When started from the command line, the IRM has the following options:
.PP
\-\-stdout
.RS 4
The \fIIRM\fR daemon and \fIIPCPs\fR will print output to stdout
instead of the system logs.
.RE

.PP
\-\-version
.RS 4
The \fIIRM\fR daemon will print the Ouroboros version to stdout and exit.
.RE

.SH IRM TOOL
The \fBirm\fR tool is used to command the Ouroboros subsystem or
individual IPCPs.

.SH WORKFLOW
To setup a \fIflow\fR between processes, Ouroboros has a number of
steps that must be taken. These steps can be taken in (almost) any
order.

.RS 4

\fBBinding a name to a program or process\fR tells the IRMd that a
certain \fIprogram\fR or \fIprocess\fR will respond to \fIflow
allocation\fR requests for the bound name. Multiple programs and
processes can be bound to the same name, and a program or process can
be bound to multiple names. This only works for programs that use the
Ouroboros API.

\fBCreating an IPC Process\fR will instantiate a new IPCP in the
system. This IPCP will be empty, with only a name and type specified,
and ready for either \fIbootstrapping\fR into a new layer or
\fIenrolling\fR in an existing layer.

\fBBootstrapping an IPC Process\fR gives it the necessary information
from the command line to start operating as a \fIlayer\fR. Each IPCP
type has its own set of configuration options.

\fBEnrolling an IPC Process\fR in a layer configures it by contacting
another IPCP that is already a member of that layer (bootstrapped or
enrolled). After \fIenrollment\fR, the IPCP is configured, and may or
may not have connections to the layer.

\fBRegistering a name in a layer\fR inserts the name in that
\fIlayer\fR's \fIdirectory\fR.  Once a name is \fIregistered\fR in a
layer, the name can be used as a destination handle for \fIflow
allocation\fR.

Ouroboros unicast IPCPs are just like any other process, and will need
to be bound and registered in a lower-rank layer before they can be
accessed by other processes.

In order to enroll an IPC process in a layer, some other member will
have to be reachable over a lower layer. IPCPs that wrap a legacy
transmission technology are all bootstrapped and thus need not enroll
as they work directly over a physical connection. Ouroboros currently
supports IPCPs over shared memory (local), L1 (raptor, experimental),
L2 (eth-llc and eth-dix) and L3 (udp). The unicast and broadcast
layers require connections to be established between IPCP components
for its operation.

\fBConnecting the management components\fR using \fImanagement
flows\fR allows management information to be sent between IPCPs so
that resources for flows can be allocated. Currently the only
component that requires management flows is the routing component.

\fBConnecting the data transfer components\fR using \fIdata transfer
flows\fR allows the routing system to forward packets between
IPCPs. This establishes adjacencies in a network of IPCPs. Components
that rely on data transfer flows are the \fIdirectory\fR and the
\fIflow allocator\fR.

.SH IRM IPCP COMMANDS
.PP
\fBirm ipcp create\fR type \fItype\fR name \fIname\fR
.RS 4
creates an IPCP process of type \fItype\fR in the system with name
\fIname\fR. Allowable values for \fItype\fR are:
.PP
\fBlocal\fR     - create a loopback IPCP.
.PP
\fBraptor\fR    - create an IPCP that attaches to a raptor NetFPGA device.
.PP
\fBeth-llc\fR   - create an IPCP that attaches to Ethernet using LLC frames.
.PP
\fBeth-dix\fR   - create an IPCP that attaches to Ethernet using DIX frames.
.PP
\fBudp\fR       - create an IPCP that attaches to a UDP socket.
.PP
\fBunicast\fR   - create a unicast IPCP that uses lower level layers.
.PP
\fBbroadcast\fR - create a broadcast IPCP that uses lower level layers.
.RE

.PP
\fBirm ipcp destroy\fR name \fIname\fR
.RS 4
destroys the IPCP process with name \fIname\fR.
.RE
.PP
\fBirm ipcp bootstrap\fR type \fItype\fR name \fIname\fR layer \fIlayer\fR \
[\fIparams\fR] [\fIautobind\fR]
.RS 4
bootstraps an IPCP process of type \fItype\fR and name \fIname\fR in a
layer \fIlayer\fR. If an IPCP with that name does not exist yet, the
IPCP will be created.
.PP
Values for [\fIparam\fR] are dependent on \fItype\fR:
.PP
\fBlocal\fR
.RS 4
.PP
[hash \fIpolicy\fR] specifies the hash function used for the directory,
.br
\fIpolicy\fR: SHA3_224, SHA3_256, SHA3_384, SHA3_512.
.br
default: SHA3_256.
.RE

\fBraptor\fR
.RS 4
.PP
[hash \fIpolicy\fR] specifies the hash function used for the directory,
.br
\fIpolicy\fR: SHA3_224, SHA3_256, SHA3_384, SHA3_512.
.br
default: SHA3_256.
.RE

.PP
\fBeth-llc\fR
.RS 4
.PP
dev \fIinterface\fR specifies the interface to bind the IPCP to.
.PP
[hash \fIpolicy\fR] specifies the hash function used for the directory,
.br
\fIpolicy\fR: SHA3_224, SHA3_256, SHA3_384, SHA3_512.
.br
default: SHA3_256.
.RE

.PP
\fBeth-dix\fR
.RS 4
.PP
dev \fIinterface\fR specifies the interface to bind the IPCP to.
.PP
[ethertype \fIethertype\fR] specifies the ethertype used for the layer.
.br
default: 0xA000.
.PP
[hash \fIpolicy\fR] specifies the hash function used for the directory,
.br
\fIpolicy\fR: SHA3_224, SHA3_256, SHA3_384, SHA3_512.
.br
default: SHA3_256.
.RE

.PP
\fBudp\fR
.RS 4
.PP
ip \fIip\fR specifies the local IP address to bind to
.PP
[dns \fdns\fR] specifies an optional DDNS server that will be used for
the directory.
.PP
[cport \fcport\fR] specifies a client UDP port that will be used for
sending packets.
.br
default: A random UDP port in the ephemeral range
.PP
[sport \fsport\fR] specifies a server UDP port that is used for
receiving ouroboros traffic. This must be the same for the entire UDP
layer. Parallel UDP layers should use different ports. This UDP port
needs to be forwarded if the server is behind a NAT.
.br
default: 3435
.RE

.PP
\fBunicast\fR
.RS 4
.PP
[addr \fIaddress_size\fR] specifies the size of an address in bytes.
.br
default: 4 bytes.
.PP
[fd \fIfd_size\fR] specifies the size of the flow descriptor field
(for identifying internal components using the DT component) in bytes.
.br
default: 2 bytes.
.PP
[ttl] specifies the maximum value for the time-to-live field.
.PP
[addr_auth \fIpolicy\fR] specifies the address authority policy.
.br
\fIpolicy\fR: flat.
.br
default: flat.
.PP
[routing \fIpolicy\fR] specifies the routing policy.
.br
\fIpolicy\fR:
.RS 4
link_state: default shortest-path link-state routing.
.br
lfa: loop-free alternate resilient routing.
.br
ecmp: equal-cost multipath routing.
.RE
.br
default: link_state.
.PP
[congestion \fIpolicy\fR] specifies the congestion avoidance policy.
.br
\fIpolicy\fR:
.RS 4
none: no congestion avoidance.
.br
mb-ecn: Multi-bit explicit congestion notification and avoidance.
.RE
.br
default: mb-ecn.
.PP
[hash \fIpolicy\fR] specifies the hash function used for the directory.
.br
\fIpolicy\fR: SHA3_224, SHA3_256, SHA3_384, SHA3_512.
.br
default: SHA3_256.
.PP
[autobind] will automatically bind a unicast or broadcast IPCP to its
name and the layer name.
.RE

.PP
\fBbroadcast\fR
.RS 4
.PP
[hash \fIpolicy\fR] specifies the hash function used for the layer name.
.br
\fIpolicy\fR: SHA3_224, SHA3_256, SHA3_384, SHA3_512.
.br
default: SHA3_256.
.RE
.RE

.PP
\fBirm ipcp enroll\fR name \fIname\fR [type \fItype\fR] [dst \fIdst\fR] \
[layer \fIlayer\fR] [\fIautobind\fR]
.RS 4
enrolls a unicast ot broadcast IPCP \fIname\fR with a dst IPCP into a
layer. Either dst or layer need to be specified.
.PP
[type \fItype\fR] specifies the IPCP type (unicast or broadcast).
.PP
[dst \fIdst\fR] specifies the destination name to enroll with
.PP
[layer \fIlayer\fR] specifies the name of the layer. Only specifying
the layer name is a shorthand for the destination name being the same
as the layer name.
.PP
[autobind] will automatically bind this IPCP to its name and the layer name.
.RE

\fBirm ipcp connect\fR name \fIname\fR component \fIcomponent\fR dst
\fIdestination\fR
.RS 4
connects a \fIcomponent\fR (\fBdt\fR or \fBmgmt\fR) of a unicast or
broadcast IPCP with name \fIname\fR to that component of the
destination IPCP within the same layer.
.RE

\fBirm ipcp disconnect\fR name \fIname\fR component \fIcomponent\fR dst
\fIdestination\fR
.RS 4
disconnects \fIcomponent\fR (\fBdt\fR or \fBmgmt\fR) of a unicast or
broadcast IPCP with name \fIname\fR from that component of the
destination IPCP within the same layer.
.RE

\fBirm ipcp list\fR type \fItype\fR name \fIname\fR layer \fIlayer\fR
.RS 4
lists IPCPs in the system. You can filter by type, by name or by layer.
.RE
.RE

.SH IRM BIND COMMANDS
.PP
\fBirm bind\fR program \fiprogram\fR name \fIname\fR [[\fIauto\fR] -- \
[\fIparam\fR] [\fIparam\fR] ... [\fIparam\fR]]
.RS 4
bind a program \fIprogram\fR, referenced by its binary, to a certain
name. The binary can be specified by an absolute or relative path, and
will be searched for in all directories specified in the \fBPATH\fR
variable. If \fIname\fR is registered in a layer, all instances of
\fIprogram\fR will accept flows for \fIname\fR. If the \fIauto\fR
keyword is specified, the IRMd will start instances if there are no
readily available instances when a new flow is requested and any
arguments \fIparam\fR that are passed after -- are passed to the
application when it is started.
.RE

.PP
\fBirm bind\fR process \fIpid\fR name \fIname\fR
.RS 4
bind a process \fIpid\fR to a certain name. If \fIname\fR is registered
in a layer, the process \fIpid\fR will accept flows for that name.
.RE

.PP
\fBirm bind\fR ipcp \fIipcp\fR name \fIname\fR
.RS 4
bind an IPCP \fIipcp\fR to a \fIname\fR. If \fIname\fR is registered
in a layer, \fIipcp\fR will accept flows for \fIname\fR. This is a
shorthand for binding the pid of \fipcp\fR.
.RE

.PP
\fBirm unbind\fR program \fIprogram\fR name \fIname\fR
.RS 4
remove the binding between \fIprogram\fR and \fIname\fR. Instances of
\fIprogram\fR will not accept future flow allocation requests for
\fIname\fR.
.RE

.PP
\fBirm unbind\fR process \fIpid\fR name \fIname\fR
.RS 4
remove the binding between \fIpid\fR and \fIname\fR. The process
\fIpid\fR will not accept future flow allocation requests for
\fIname\fR.
.RE

.PP
\fBirm unbind\fR ipcp \fIipcp\fR name \fIname\fR
.RS 4
remove the binding between \fIipcp\fR and \fIname\fR. This IPCP will
not accept future flow allocation requests for \fIname\fR.
.RE

.SH IRM NAME COMMANDS
.PP
\fBirm name create \fIname\fR \fIlb\fR policy
.RS 4
Create a name \fIname\fR with a load-balancing policy
.br
\fIpolicy\fR: round-robin, spillover
.br
.RE

.PP
\fBirm name destroy \fIname\fR
.RS 4
Destroy name \fIname\fR. This does not unregister or unbind it.
.RE

.PP
\fBirm name register \fIname\fR \fIipcp\fR ipcp [\fIipcp\fR ...]
layer [layer \fIlayer\fR ...]
.RS 4
Register name \fIname\fR in ipcps \fIipcp\fR ipcp and layers \fIlayer\fR.
.RE

.PP
\fBirm name unregister \fIname\fR \fIipcp\fR ipcp [\fIipcp\fR ...]
layer [layer \fIlayer\fR ...]
.RS 4
Unregister name \fIname\fR in ipcps \fIipcp\fR ipcp and layers \fIlayer\fR.
.RE

.PP
\fBirm name list \fIname\fR
.RS 4
List names in the system. \fIname\fR can be used as a prefix to filter
the names.
.RE


.SH TERMINOLOGY
Please see \fBouroboros-glossary\fR(7).

.SH AUTHORS
Ouroboros was started and is currently maintained by Dimitri Staessens
and Sander Vrijders.

.SH REPORTING BUGS
Report bugs on the bugzilla tracker at https://ouroboros.rocks/bugzilla

.SH SEE ALSO
.BR flow_alloc "(3), " ouroboros-tutorial "(7), " ouroboros-glossary (7)

.SH COLOPHON
This page is part of the Ouroboros project, found at
http://ouroboros.rocks

These man pages are licensed under the Creative Commons Attribution
4.0 International License. To view a copy of this license, visit
http://creativecommons.org/licenses/by/4.0/