Version 1 (modified by Aaron Helsinger, 9 years ago) (diff)


GENI Aggregate Manager API Version 3 Common Concepts

This page documents the GENI Aggregate Manager API version 3 common concepts.

The current officially adopted version of the API is 2 and is documented on the main API page.

The current adopted or proposed changes for the AM API, whose impacts are reflected here, are documented on the AM API Draft wiki page.

For a summary of the method signatures of the Aggregate Manager API if all changes are adopted, see the draft Method Signatures page.

Common Arguments, Returns, and Concepts

structs and optional arguments

Unless otherwise specified, all arguments and returns of type struct may include aggregate or resource-specific entries. As arguments, such options must be optional for the client to supply, with the aggregate providing a reasonable default.

RSpec data type

Throughout this API, multiple arguments and returns are labeled as an RSpec. These fields shall be understood as XML documents following one of the schemas advertised in the return from GetVersion. All such RSpecs must pass an XML schema validator, must list all used schemas and namespaces within the document, using schemas that are publicly available. The GetVersion return advertises schemas for advertisment and request RSpecs; the schemas for manifest RSpecs are assumed to be available at the same base URL, but using a corresponding manifest schema.

A fully GENI AM API compliant aggregate will always support the GENI standard schemas for RSpecs, available at As of 4/2012, the current GENI RSpec version is 3. Aggregates are free to use an alternate format internally, but must accept and produce compliant RSpecs on demand.

More information on GENI RSpecs is available on the ProtoGENI wiki.

The Aggregate Manager API requires this contract: Aggregates advertise the type and version of RSpec formats that they support. If available, they specify the schema, namespace and extensions combination which is the authoritative definition of that format. Clients of the API should understand that combination in order to know how to understand the resources available at that aggregate.

If an aggregate advertises a particular type/version (optionally defined with a combination of schema, namespace and extensions) in the geni_ad_rspec_versions attribute of GetVersion, then it promises to send a correct Advertisement RSpec in response to a ListResources call which supplies a geni_rspec_version option containing that type/version. (geni_rspec_version is a struct with 2 members, type and version. type and version are case-insensitive strings, matching those in geni_ad_rspec_versions).

If an Aggregate advertises a particular type/version (optionally defined with a combination of schema, namespace and extensions) in the geni_request_rspec_versions attribute of GetVersion then it promises to correctly honor an Allocate (was CreateSliver in API v2) call containing a request RSpec in the given format, and then to return a Manifest RSpec in the corresponding format (i.e. a GENI format request is answered with a GENI format manifest). The aggregate also promises to send a correctly formatted Manifest RSpec in response to a Describe call which supplies a valid slice URN or list of sliver URNs and an geni_rspec_version option containing that supported type/version.

In this API, such RSpec fields are labeled as type geni.rspec.


Many methods take an array of credentials. This is an array of credential type, version, and string value.

credentials = [
    geni_type: <string>, 
    geni_version: <string>,
    geni_value: <string>,

Each credential (in geni_value) is defined as a signed document. A given list of credentials may contain credentials in multiple formats. The list may be empty. A given authorization policy at an AM may require 0, 1, or many credentials. Aggregates are required to allow credentials which are not used by local authorization policy or engines, using only credentials locally relevant.

  • An AM must pick credentials out of the list that it understands and be robust to receiving credentials it does not understand.
  • Current slice and user credentials will be recognizable for following the schema defined in Change Set K and GeniApiCredentials.
  • AMs are required to continue to accept current-format credentials.
    • In particular, a single standard slice credential remains sufficient for most authorization policies.
  • Other credential formats acceptable by some aggregates might include ABAC x509 Attribute certificates, eg.
  • AMs may get other authorization material from other sources: EG a future Credential Store service.

At least one subset of the credentials (e.g. a single SFA style slice credential) must authorize operations for the slice specified in slice_urn if that is an argument, or for the slice that contains the named slivers, if sliver urns are an argument, or a valid set of administrative credentials with sufficient privileges. When sliver_urns are supplied, all such slivers must belong to the same slice, over which the given credential set provides access. Credentials must be valid (signed by a valid GENI certificate authority either directly or by chain, not expired, and grant privileges to the client identified by the SSL client certificate). Each method requires specific privileges, which must be granted by the provided credentials. Note that the semantics of this argument is not clear: most implementations require a single credential to provide all needed privileges. Alternative interpretations might, for example, accumulate privileges from each valid credential to determine overall caller permissions. For details on GENI AM API format credentials, see the GENI wiki.


An XML-RPC struct. For GetVersion only, this argument is optional. In all other methods, it is required. Only ListResources has required entries in the options struct. Aggregates are compliant with this API change by accepting this argument. Aggregates may accept entries to this struct. Aggregates should not require any new options to any method - they should always have a reasonable default for any such option. Aggregates should document new options arguments. The prefix geni_ is reserved for members that are part of this API specification. Implementations should choose an appropriate prefix to avoid conflicts.

Operations on Individual Slivers

A Sliver is an aggregate defined grouping of resources within a slice at this aggregate, whose URN identifies the sliver, and can be used as an argument to methods such as Delete or Renew, and whose status can be independently reported in the return from Status. The AM defines 1 or more of these groupings to satisfy a given resource request for a slice. All reserved resources are directly contained by exactly 1 such sliver container, which is in precisely 1 slice.

One or more slivers are created by an aggregate when the experimenter calls Allocate(). This API encourages aggregates to independently manage each sliver, allowing experimenters to selectively Delete, Renew, or Provision each sliver. As such, these methods take a list of sliver urns (or a slice urn), and return a struct reporting results for each sliver URN independently. However, slivers at an aggregate may have interdependencies, and an individual aggregate may not be able to independently manage each sliver, without also modifying other related slivers. This API defines a number of aggregate configuration options returned by GetVersion, and an option to many methods, allowing aggregates to advertise their behavior, and experimenters to request particular behavior.

  1. geni_single_allocation: <XML-RPC boolean 1/0, default 0>: When performing one of (Describe, Allocate, Renew, Provision, Delete), the AM requires you to include either the slice urn or the urn of all the slivers in the same state. If you attempt to run one of those operations on just some slivers in a given state, the AM will return an error.

For example, you must Provision all geni_allocated slivers at once: At an aggregate with geni_single_allocation true, if you supply a list of sliver URNs to Provision that is only 'some' of the geni_allocated slivers for this slice at this AM, then the AM will return an error. Similarly, such an aggregate would return an error from Describe if you request a set of sliver URNs that is only some of the geni_provisioned slivers.

  1. geni_allocate: A string, one of fixed set of possible values. Default is geni_single. This option defines whether this AM allows adding slivers to slices at an AM (i.e. calling Allocate() multiple times, without first deleting the allocated slivers). Possible values:
    • geni_single: Performing multiple Allocates without a delete is an error condition because the aggregate only supports a single sliver per slice or does not allow incrementally adding new slivers. This is the AM API v2 behavior.
    • geni_disjoint: Additional calls to Allocate must be disjoint from slivers allocated with previous calls (no references or dependencies on existing slivers). The topologies must be disjoint in that there can be no connection or other reference from one topology to the other.
    • geni_many: Multiple slivers can exist and be incrementally added, including those which connect or overlap in some way. New aggregates should strive for this capability.

Many methods also take a new option (aggregates must support it, clients do not need to supply it):

   geni_best_effort: <XML-RPC boolean 1/0, default 0>

If false, the client is requesting that the aggregate either fully satisfy the request, moving all listed slivers to the desired state, or fully fail the request, leaving all slivers in their original state. If the aggregate cannot guarantee all or nothing success or failure given the included slivers and resource types, the aggregate shall fail the request, returning an appropriate error code. If this option is true, then some slivers may transition to the new state, and some note. Aggregates must examine the return closely to know the state of their slivers.

It is expected that many aggregates will implement one of the following combinations of options:

  • Accept requests for geni_best_effort = true, and advertise geni_allocate = geni_many, geni_single_allocation = false (E.G. FOAM, PlanetLab).
  • Operate as though all requests were geni_best_effort = false, and advertise geni_allocate = geni_disjoint, geni_single_allocation = true (E.G. ProtoGENI).


Several methods take some URNs to identify what to operate on. These methods are defined as accepting a list of arbitrary strings we call URNs. This API defines two kinds of URNs that may be supplied here, slice URNs and sliver URNs (see the GENI identifiers page, and Change Set K). Some aggregates may understand other URNs, but these are not defined or required here. Aggregates that accept only URNs defined by this API will return an error when given URNs not in one of those forms. This API requires that aggregates accept either a single slice URN, or 1+ sliver URNs that all belong to the same slice. Aggregates are not required to accept both a slice URN and sliver URNs, 2+ slice URNs, or a set of sliver URNs that crosses multiple slices. Some aggregates may choose to accept other such combinations of URNs. Aggregates that accept only arguments defined by this API will return an error when given more than 1 slice URN, a combination of both slice and sliver URNs, or a set of sliver URNs that belong to more than 1 slice.

Sliver Allocation States

Many operations in this API create slivers or change the allocation status of slivers, and often return the current allocation status of each sliver. Valid sliver allocation states are:

  1. geni_unallocated (alternatively called 'null'). The sliver does not exist. This is the small black circle in typical state diagrams.
  2. geni_allocated (alternatively called 'offered' or 'promised'). The sliver exists, defines particular resources, and is in a sliver. The aggregate has not (if possible) done any time consuming or expensive work to instantiate the resources, provision them, or make it difficult to revert the slice to the state prior to allocating this sliver. This state is what the aggregate is offering the experimenter.
  3. geni_provisioned. The aggregate has started instantiating resources, and otherwise making changes to resources and the slice to make the resources available to the experimenter. At this point, operational states are valid to specify further when the resources are available for experimenter use.

geni_allocated represents resources that have been allocated to a slice without provisioning the resources. This represents a cheap and un-doable resource allocation. When a sliver is created and moved into state 2 (geni_allocated), the aggregate produces a manifest RSpec identifying which resources are included in the sliver. These resources are exclusively available to the containing sliver, but are not ready for use. In particular, allocating a sliver should be a cheap and quick operation, which the aggregate can readily un-do without impacting the state of slivers which are fully provisioned. For some aggregates, transitioning to this state may be a no-op.

States 2 and 3 (geni_allocated and geni_provisioned) have aggregate and possibly resource specific timeouts. By convention the geni_allocated state timeout is typically short. The geni_provisioned state timeout is the sliver expiration. If the client does not transition the sliver from geni_allocated to geni_provisioned before the end of the geni_allocated state timeout, the sliver reverts to geni_unallocated. If the experimenter needs more time, the experimenter should be allowed to request a renewal of either timeout. Note that typically the sliver expiration time (timeout for state 3, geni_provisioned) will be notably longer than the timeout for state 2, geni_allocated.

State 3, geni_provisioned, is the state of the sliver allocation after the aggregate begins to instantiate the sliver. Note that fully provisioning a sliver may take noticeable time. This state also includes a timeout - the sliver expiration time (which is not necessarily related to the time it takes to provision a resource). Renew extends this timeout. For some aggregates and resource types, moving to this state from state 2 (geni_allocated) may be a no-op.

If the transition from one state to another fails, the sliver shall remain in its original state.

  1. Allocate moves 1+ slivers from geni_unallocated (state 1) to geni_allocated (state 2). This method can be described as creating an instance of the state machine for each sliver. If the aggregate cannot fully satisfy the request, the whole request fails. This is a change from the version 2 CreateSliver, which also provisioned the resources, and 'started' them. That is Allocate does 1 of the 3 things that CreateSliver did previously.
  2. Delete moves 1+ slivers from either state 2 or 3 (geni_allocated or geni_provisioned), back to state 1 (geni_unallocated). This is similar to the AM API version 2 DeleteSliver.
  3. Renew, when given allocated slivers, requests an extended timeout for slivers in state 2 (geni_allocated).
  4. Renew can also be used to request an extended timeout for slivers in state 3 - the geni_provisioned state. That is, this method's semantics can be the same as RenewSliver from AM API v2.
  5. Provision moves 1+ slivers from state 2 (geni_allocated) to state 3 (geni_provisioned). This is some of what version 2 CreateSliver did. Note however that this does not 'start' the resources, or otherwise change their operational state. This method only fully instantiates the resources in the slice. This may be a no-op for some aggregates or resources.

These states apply to each sliver individually. Logically, the state transition methods then take a single sliver URN. For convenience, these methods accept a list of sliver URNs, or a slice URN as a simple alias for all slivers in this slice at this aggregate.

FIXME: Add picture

Sliver Operational States

The AM API defines a few operational states with particular semantics. AMs are not required to support them for a given set of resources, but if they use them, they must follow the given semantics. AMs are however STRONGLY encouraged to support them, to provide maximum utility. There is one state that AMs are required to support, geni_pending_allocation, for a sliver which has not been fully allocated and provisioned.

AMs may have their own operational states/state-machine internally. AMs are required to advertise such states and actions that experimenters may see or use, by using Ad RSpec extensions. Operational states which the experimenter never sees, need not be advertised. Operational states and actions are generally by resource type. The standard RSpec extension attaches such definitions to the sliver_type element of RSpecs.

States should be defined in terms of (a) whether the resource is accessible to the experimenter on the data or control planes, (b) whether an experimenter action is required to change from this state, and if so, (c) what action or actions are useful. If the resource will change states without explicit experimenter action, what is the expected next state on success.

Note that states represent the AM's view of the operational condition of the resource. This state represents what the AM has done or learned about the resource, but experimenter actions may cause failures that the AM does not know about.

There is no busy state. Instead, AMs are encouraged to define separate such transition states for each separate transition path, allowing experimenters to distinguish the start and end states for this transition.

shutdown is not an operational state for a sliver. The Shutdown() API method applies to an entire slice.

Operational states are generally only valid for slivers which have been provisioned (geni_provisioned allocation state).

GENI defined operational states:

  • geni_pending_allocation: A wait state. The sliver is still being allocated and provisioned, and other operational states are not yet valid. PerformOperationalAction may not yet be called on this sliver. For example, the sliver is in allocation state geni_provisioned, but has not been fully provisioned (e.g., the VM has not been fully imaged). Once the sliver has been fully allocated, the AM will transition the sliver to some other valid operational state, as specified by the advertised operational state machine. Common next states are geni_notready, geni_ready, and geni_failed.
  • geni_notready: A final state. The resource is not usable / accessible by the experimenter, and requires explicit experimenter action before it is usable/accessible by the experimenter. For some resources, geni_start will move the resource out of this state and towards geni_ready.
  • geni_configuring: A wait state. The resource is in process of changing to geni_ready, and on success will do so without additional experimenter action. For example, the resource may be powering on.
  • geni_stopping: A wait state. The resource is in process of changing to geni_notready, and on success will do so without additional experimenter action. For example, the resource may be powering off.
  • geni_ready: A final state. The resource is usable/accessible by the experimenter, and ready for slice operations.
  • geni_ready_busy: A wait state. The resource is performing some operational action, but remains accessible/usable by the experimenter. Upon completion of the action, the resource will return to geni_ready.
  • geni_failed: A final state. Some operational action failed, rendering the resource unusable. An administrator action, undefined by this API, may be required to return the resource to another operational state.

Sliver Operational Actions

The API defines a few operational actions: these need not be supported. AMs are encouraged to support these if possible, but only if they can be supported following the defined semantics.

AMs may have their own operational states/state-machine internally. AMs are required to advertise such states and actions that experimenters may see or use, by using Ad RSpec extensions. Operational states which the experimenter never sees, need not be advertised. Operational states and actions are generally by resource type. The standard RSpec extension attaches such definitions to the sliver_type element of RSpecs.

Tools must use the operational states and actions advertisement to determine what operational actions to offer to experimenters, and what actions to perform for the experimenter. Tools may choose to offer actions which the tool does not understand, relying on the experimenter to understand the meaning of the new action.

Any operational action may fail. When this happens, the API method should return an error code. The sliver may remain in the original state. In some cases, the sliver may transition to the geni_failed state.

GENI defined operational actions:

  • geni_start: This action results in the sliver becoming geni_ready eventually. The operation may fail (move to geni_failed), or move through some number of transition states. See EG booting a VM.
  • geni_restart: This action results in the sliver becoming geni_ready eventually. The operation may fail (move to geni_failed), or move through some number of transition states. During this operation, the resource may or may not remain accessible. Dynamic state associated with this resource may be lost by performing this operation. See EG re-booting a VM.
  • geni_stop: This action results in the sliver becoming geni_notready eventually. The operation may fail (move to geni_failed), or move through some number of transition states. See EG powering down a VM.

Return Struct

code, value, and output together provide the standard return from all AM API methods.

A struct indicating the success or failure of this call at the Aggregate Manager. It consists of 1 required field and 2 optional fields.
  struct code = {
       int geni_code;
       [optional: string am_type;]
       [optional: int am_code;]
Method-specific. Required on success. Optional on error.
On failure or error, this is required. Optional on success. This is an XML-RPC string with a human readable message explaining the result. Specifically, this might include an error string, a stacktrace, or other useful messages to help the experimenter resolve or report the failure or error. It is not defined on success, though aggregates are free to use it.

Implementations can add additional members to the return struct as desired. The prefix geni_ is reserved for members that are part of this API specification. Implementations should choose an appropriate prefix to avoid conflicts. Aggregates should document any additional return values.

Aggregates shall return consistent values for geni_code as described here. Aggregates wishing to be more specific may use the am_type and am_code values.

Success is always indicated using a geni_code value of 0.

On one of the error or failure cases listed in the table below, aggregates shall return the indicated error code.

Elements in code

An integer supplying the GENI standard return code indicating the success or failure of this call. Error codes are standardized and defined in the attached XML document. Codes may be negative. A success return is defined as geni_code of 0.
Optional. A (case insensitive) string indicating the type of Aggregate Manager running locally. When an aggregate wants to return an aggregate specific return code in the am_code field, they supply an am_type to qualify the kind of aggregate specific return code they are supplying. This is the namespace of the aggregate specific return code. This field is optional: aggregates are not required to supply an aggregate specific return code, and clients need not look at it. This code further qualifies the kind of error or success that the aggregate is returning, as primarily defined by the value of geni_code. Standard values for am_type are defined in the attached XML document.
An integer supplying the more specific return code, relative to the aggregate type specified in am_type. This integer may be negative. Aggregates should document these codes publicly. This API does not specify how or where that documentation should be provided.

Aggregates are encouraged to use code values and output messages that help experimenters and tools distinguish between bad input, other experimenter error, temporary server errors, or server bugs.

GENI standard error codes are documented in the attached XML document, and listed below.

0 SUCCESS "Success"
1 BADARGS "Bad Arguments: malformed arguments"
2 ERROR "Error (other)"
3 FORBIDDEN "Operation Forbidden: eg supplied credentials do not provide sufficient privileges (on given slice)"
4 BADVERSION "Bad Version (eg of RSpec)"
5 SERVERERROR "Server Error"
6 TOOBIG "Too Big (eg request RSpec)"
7 REFUSED "Operation Refused"
8 TIMEDOUT "Operation Timed Out"
9 DBERROR "Database Error"
11 UNAVAILABLE "Unavailable (eg server in lockdown)"
12 SEARCHFAILED "Search Failed (eg for slice)"
13 UNSUPPORTED "Operation Unsupported"
14 BUSY "Busy (resource, slice); try again later"
15 EXPIRED "Expired (eg slice)"
16 INPROGRESS "In Progress"
17 ALREADYEXISTS "Already Exists (eg the slice}"

Aggregates are similarly encouraged to provide hints on how to fix bad requests using the value entry to experimenters on error or failures. For example, a failed RenewSliver call that failed because you are not allowed to renew your sliver that far in the future, might return a new date string in the value field that would be allowed. Similarly, a failed CreateSliver call might return a modified request RSpec in the value field.

Note that a malformed XML-RPC request should still raise an XML-RPC Fault, and other Faults dictated by the XML-RPC specification should still be raised. Aggregates should avoid raising an error (XML-RPC Fault) for application layer errors or any other cases where the XML-RPC specification does not require a Fault, but rather should attempt to return this struct, providing any error messages and stack traces in the output field or other additional fields. Certain XML-RPC errors may be returned using Faults or otherwise by the XML-RPC layer, or may more properly be returned using this struct in the application layer. In such cases, servers should use error codes with negative values. Selected such errors are listed below:

-32001 SERVERBUSY "Server is (temporarily) too busy; try again later"

Note also that servers may respond with other HTTP error codes, and clients must be prepared to deal with those situations. Specifically, a server that is busy might return HTTP code 503, or just refuse the connection.

datetime data type

All datetime arguments and returns in this API shall conform to RFC 3339. This represents a subset of the valid date/time strings permissible by the standard XML-RPC date/time data type, dateTime.iso8601.

  • Full date and time with explicit timezone: offset from UTC or in UTC)
  • e.g.: 1985-04-12T23:20:50.52Z or 1996-12-19T16:39:57-08:00

In the specification of this API, this is described as dateTime.rfc3339.


The geni_end_time argument requests an expiration of the specified slivers. It is in dateTime.rfc3339 format (defined above). When an explicit argument, it is required, and aggregates must honor the request to the extent local policy permits. When an option in the options struct, clients may omit the option, and AMs may choose not to or be unable to honor this option, but may still succeed the overall request.


geni_best_effort: <XML-RPC boolean 1 or 0, default false (0)>

Clients may omit this option, but aggregates must honor the option if possible. This option modifies the way that the operation applies to all named slivers. By default (geni_best_effort=false), the operation must apply equally to all slivers, either succeeding or failing for all. When true, the aggregate may succeed the operation for some slivers, while failing the operation for other slivers.


struct geni_users[] is an option for some methods.

Clients may omit this option. Aggregates should honor this option for any resource that accepts the provided login keys, and ignore it for other resources. This option is an array of user structs, which contain information about the users that might login to the sliver that the AM needs to know about. For example, this option is the mechanism by which users supply their SSH public keys, permitting SSH login to allocated nodes. In such cases, the corresponding manifest RSpec will contain the ssh-users element on each such node, showing the login username and applicable public keys. When this option is supplied, each struct must include the key 'keys', which is an array of strings and can be empty. The struct must also include the key 'urn', which is the user’s URN string. For example:

    keys: [<ssh key>, <ssh key>]
    keys: [<ssh key>]

Documenting Aggregate Additions

Aggregates are free to add additional return values or input options to support aggregate or resource specific functionality, or to innovate within the bounds of the AM API. This includes adding new methods that use the same transport, interface, certificates, and credentials. Aggregates are encouraged to document any such new return values which they return or options arguments, to bootstrap coordination with clients, and provide documentation for human experimenters. One way to provide partial documentation, is to implement XML-RPC introspection. Through the use of method help, aggregates can provide human readable text describing return values. Alternatively or additionally, aggregates may document return values as part of their return from GetVersion. This API does not specify the format for advertising those extra return values in GetVersion.

Supporting Multiple API Versions

Aggregates are free to support multiple versions of the AM API. They do so by providing different URLs for each version of the API that they support. Aggregates should have a 'default' URL (the one typically advertised). That version runs whichever version of the API the server chooses (could be the latest, could be something else.)

When aggregates start supporting a new version of the API, they should keep running the old version of the API for a suitable transition period.

Aggregates running multiple versions of the API must advertise the URLs and versions of the API supported using the new GetVersion return as part of the value entry:

geni_api_versions: an XML-RPC struct containing 1+ entries of:
  Name: Integer - supported GENI AM API version
  Value: String - URL to the XML-RPC server implementing that version of the GENI AM API

For example

geni_api_versions: {
  1: <URL>,
  2: <Local URL, as this is API version 2>,

The entries indicate versions of the API that are supported, and URLs are absolute URLs where that version of the API is supported.

Attachments (3)

Download all attachments as: .zip