GENI identifies objects (be it a researcher, resource, clearinghouse, or aggregate manager) with a Uniform Resource Name (URN).
The format of a GENI URN is: urn:publicid:IDN+<authority string>+<type>+<name>
The GENI URN format is adapted from the
ProtoGENI URN format
and the
GMOC GENI URN proposal. All sections of the URN are mandatory. Note that additional '+' characters are allowed in the <name>
section.
GENI URNs are in the URN namespace for Public Identifiers. As such, each GENI URN is of the form: urn:publicid:{transcribed-public-identifier}
. RFC 3151 describes how public identifiers are transcribed to URNs (which involves collapsing whitespace and replacing certain characters with %
encoded values).
The public identifier section of a GENI URN must begin with "IDN ". When transcribed, this means that all GENI URNs begin with "urn:publicid:IDN+".
To transcribe a public ID to a URN use the following rules, from RFC 3151:
From | Transcribe to |
---|---|
leading and trailing whitespace |
trim |
whitespace |
collapse to a single '+' |
'//' |
':' |
'::' |
';' |
'+' |
'%2B' |
":' |
'%3A' |
'/' |
'%2F' |
';' |
'%3B' |
'!'' |
'%27' |
'?' |
'%3F' |
'#' |
'%23' |
'%' |
'%25 |
Example translations:
Public ID | GENI URN |
---|---|
IDN plc//princeton authority sa |
urn:publicid:IDN+plc:princeton+authority+sa |
IDN gcf//gpo//gpolab user joe |
urn:publicid:IDN+gcf:gpo:gpolab+user+joe |
IDN gcf//gpo//gpolab node switch 1 port 2 |
urn:publicid:IDN+gcf:gpo:gpolab+node+switch+1+port+2 |
Formally, the public ID format follows the following naming convention (in perl regular expression syntax for the encoding string):
IDN [toplevelauthority][\/\/sub-authority]* resource-type resource-name
This would lead to the following URN schema in the public id namespace (Using the transformation in RFC 3151 :
urn:publicid:IDN+toplevelauthority[:sub-authority]*\+resource-type\+resource-name
Where:
- toplevelauthority
-
is unique across the GENI federation and tied to a specific authority instance. By convention it should be an internationalized domain name (which must then match the one in the certificate of the authority which issued the object name)
- sub-authority
-
is unique within the toplevelauthority and by convention is often an internationalized domain name (which should then match the one in the certificate of the authority which issued the object name)
- resource-type
-
is a string describing the type of the named object (the set of strings is described [#Type below])
- resource-name
-
should uniquely identify the object among any other resources with identical toplevelauthority+sub-authority and resource-type. It is important to realize that the GENI API attaches no other significance to this field, and in particular, no relation is implied between objects with identical resource-name but differing toplevelauthority or resource-type. However, individual authorities may choose to define additional semantics for resource names.
The following examples are borrowed from the GMOC Proposal:
Resource | GENI Identifier |
---|---|
User |
|
Planetlab node: |
|
Interface |
|
Examples of additional types of objects from the ProtoGENI documentation:
Resource | GENI Identifier |
---|---|
Slice |
|
The Utah Emulab slice authority |
|
Sliver |
|
In the GENI API, URNs are used to name slices (as seen as arguments in GENI API calls), to identify users, and to label resources. URNs are also used in GENI certificates (to bind public keys to identifiers) and in Credentials (to grant permissions to source identifiers on target identifiers).
Authorities are resource namespaces. For instance, 'plc' is the overall
PlanetLab namespace and 'plc.princeton' and 'plc.bbn' are specific
namespaces for the Princeton and BBN PlanetLab sites. The authority
string format is toplevelauthority:sub_authority1:…:sub_authority_n
.
Only entities with URNs of type authority
are allowed to sign credentials
for a namespace (except in the case of delegation).
For example, a ProtoGENI Clearinghouse with URN like protogeni.utah… cannot issue a slice credential giving a user privileges on a PlanetLab slice (with urn plc.princeton…). Only PlanetLab can grant rights over PlanetLab slices.
Essentially, to be a valid authority for a namespace, the authority name must
be a prefix of the names in its namespace. For example, a\.b
is an authority
for, a\.b.c.d
, but a
is not an authority for, a\.b.c.d
(the subject’s
name starts with a.b
, where we’ve escaped the .
). Also any authority name
is an authority for itself.
Authority strings are to be treated as case insensitive for comparison purposes, though they should be case sensitive for display purposes (i.e. case preserving). This is consistent with the DNS standard as these are generally hostnames.
The authority in a URN should follow certain conventions for different URN types. The authority section of a URN issued by an authority must match the authority section of the URN of the authority that issued the URN, which is the authority that manages the existence of these objects.
-
slice URNs are issued by a slice authority ("+authority+sa"), so the authority section of the slice URN matches that of the slice authority
-
user URNs are issued by a member authority or slice authority or clearinghouse ("+authority+ma", "+authority+sa", "+authority+ch")
-
sliver URNs are issued by an aggregate ("+authority+am")
-
component URNs ("node" or "link" types for example) are issued by an aggregate manager or a component manager ("+authority+am" or "+authority+cm")
The <type> string has not yet been completely specified, but is used to identify the kind of resource being identified. Many APIs require use of URNs with particular types. Use the defined types below, unless they simply do not fit for you. In which case, please contact the GPO to inform them of your new type.
- authority
-
A GENI service (i.e. an XMLRPC server). By convention,
resource-name
isam
for an aggregate manager,sa
for a slice authority. In principle, other names could be used for authorities. - interface
-
A component which is an interface (an endpoint for links). The
toplevelauthority
must match that of the aggregate manager (or component manager). - link
-
A component which is a network link (a connection between two or more interfaces). The
toplevelauthority
must match that of the aggregate manager. - node
-
A component which is a node (an abstraction for networkable resources). The
toplevelauthority
must match that of the aggregate manager. - slice
-
A resource container. Every valid sliver belongs to exactly one slice. The
toplevelauthority
must match that of the slice authority. - sliver
-
A collection of resources which has been instantiated within a slice at a single aggregate. The
toplevelauthority
must match that of the aggregate manager; the corresponding slice is not identified in the sliver URN. - user
-
A GENI end user. Users are associated with slice or member authorities, but not unique slices; the slice to user correspondence is potentially many-to-many. The
toplevelauthority
must match that of the issuing authority. - vlan
-
A shared VLAN, that may cross slices. This VLAN may or may not cross aggregates, as in the GENI openflow mesoscale. VLAN URNs should name the VLAN, not identify the specific VLAN tag, as that may vary across aggregates and switches.
- tool
-
A GENI tool (i.e. a hosted experimenter tool like the GENI Portal). The
toplevelauthority
indicates the authority certifying that the holder of a certificate naming this URN operates a tool with the given name.
Additional types may be defined in the future.
Note: For ProtoGENI compatibility, slice credentials should be signed by an authority with name "sa", e.g., urn:publicid:IDN+gcf:gpo+authority+sa.
The <name> string can in general be any valid string constrained by the URN character set (e.g. no whitespace).
However, names for certain URN types are restricted:
- Slice
-
-
Slice URN alone is a label - unique at a point in time but not over time.
-
Format:
urn:publicid:IDN+<SA name>+slice+<slice name>
-
Slice names are ⇐19 characters, only alphanumeric plus hyphen (no hyphen in first character):
'^[a-zA-Z0-9][-a-zA-Z0-9]\{0,18\}$'
-
Slice names are case insensitive for comparison purposes, but should be treated as case sensitive for display purposes. In other words, servers should preserve the case.
-
- Sliver
-
-
Sliver URN should be unique over time within an AM for record-keeping / operations purposes.
-
Format:
urn:publicid:IDN+<AM name>+sliver+<sliver name>
-
Sliver names
-
Must be unique over time within that AM, and are selected by the AM.
-
May use only alphanumeric characters plus hyphen, period, or underscore:
'[a-zA-Z0-9\.\-_]+'
-
-
- User
-
-
Usernames are case-insensitive internally, though they may be case-sensitive in display.
-
For example, JohnSmth as a display name is johnsmth internally, and there cannot also be a user
JOHNSMTH
. -
In other words, usernames are case insensitive for comparison but case preserving.
-
-
Usernames should begin with a letter and be alphanumeric or underscores; no hyphen or '.':
('^[a-zA-Z][\w]\{0,7\}$')
. -
Usernames are limited to 8 characters.
-
User URNs (which contain the authority name and the username) are required to be temporally and globally unique.
-
- Tool
-
-
Tool names are case-insensitive internally, though they may be case-sensitive in display.
-
For example, JohnSmth as a display name is johnsmth internally, and there cannot also be a user
JOHNSMTH
.
-
-
Tool names should begin with a letter and be alphanumeric or underscore, hyphen, at sign or period:
('^[a-zA-Z][a-ZA-Z0-9\-_@\.]{0,63}$')
. -
Tool names are limited to 64 characters.
-
Tool URNs (which contain the authority name and the tool instance name) are required to be temporally and globally unique.
-
Tool names should encode both the tool type and instance. For example
portal-gpo
orgenidesktop-uky
.
-