Often there will be a relationship of the following kind between two keys F and N. We present two alternative descriptions:

The result of invoking key F depends on the parameters passed on the most recent invocation of key N.
This description, being operational, has the benefit that it does not describe any more than desired, in that it does not imply a specific implementation.

A more intuitive way to describe such a relationship is to say that keys F and N each point to, or designate, a common object with the ability to store parameters, and that F and N contain some other component that distinguishes them. For instance, F may be a fetch key and N a node key to the same node.
This description is generally much easier to understand, so we generally use it. The implied implementation is almost always the actual implementation, and that information gives clues to the costs of the key that are often otherwise unstated. One should keep in mind, though, that it is always possible and often useful to view a key as atomic.

The first, and most important, is the key type, or authority. The key type defines what actions the holder of the key may exercise.

The second component may be thought of as the value of the key. For keys which designate a page or node, this value contains information which the kernel can use to locate the designated object.

The third component is called the (_databyte). The databyte modifies the interpretation of some types of keys.

The databyte contains eight bits of information which may be set in at least the ways described at (p2, nodemake) and (p2, crent).

Key Taxonomy

The following list describes the sixteen types of keys which are known to the kernel. Every key is just one of these types. (p2,primary) defines the effect of jumping to each type of key.
Eight Types of Keys that Designate Nodes
A (_Node Key) provides the holder the maximum authority over the node that it designates. Keys can be copied into and retrieved from arbitrary slots in the node with a node key. See (p2,node). There is no direct way to tell if there is a process in the node nor to change that state.

A (_Fetch Key) is like a node key, except keys cannot be placed in the node, only retrieved. See (p2,fetch).

A (_Sense Key) is like a fetch key except that the operation of fetching a key from the node is accompanied by an attenuation of the fetched key. See (p2,sense).

A (_Domain Key) provides the holder with the authority to intervene in the execution by the domain of its program. See (p2,domkey).

s and (_Resume Key)s provide authority to send a message to the domain built in part from the designated node. See (p1,gates). There are three variants of resume key: restart, fault and return. The type of each of these is "Resume".

A (_Segment Key) provides authority to access the segment {(p1,segment)} defined by the node {(p2,memkeydm)}. See (p1,memtree).

A (_Meter Key) provides authority to consume the resources defined by the contents of the slots of the node. See (p1,meter).

Keys That Designate Pages
A (_Page Key) provides the authority to access the data in a page. See (p1,memtree) and (p2,pagekey).

The (_Data Key) is a construct that is a key but directly holds a non-negative integer. The data key provides no authority. We use "DK(n)" to designate a data key that holds the value n. In this implementation the value is less that 2**56. The value will always be less than 2**128.
The value in a data key is frequently interpreted or produced by the kernel. In these situations the value is interpreted in binary and the bits numbered by negative integers. The unit's position is -1, the two's position is -2, the four's position is -3, etc. Older parts of the manual number the same bits 47, 46, 45, etc. This numbering was invented when it was thought that there would never be more than 48 bits.

See (p2,freadat) for specifications.

{arcane}A (_Node Range Key) or (_Page Range Key) is a closely held key which provides the holder with strong authority over a group of nodes or pages. It is used to allocate space on disks. See (p2,rangekey) and (p2,bank).

are several individual keys that designate several one-of-a-kind objects that are implemented by the kernel.

Keys which will be generally available are described individually in (p2,discrim), (p2,datac), (p2,retner), (p2,metertran). Other keys will be closely held by trusted domains. These closely held keys are described in (p2,resprim).

is a key that may be used to glean information from the kernel concerning the use of real storage by selected users. See (p2,soon-sets).

is a rescindable key to an I/O device. Stylized channel programs may be specified for the device by the holder of this key. See (p2,devicekey).

See (p2,spage) about a proposed extension and key type.

Other Classifications of Keys
The following are some classifications of keys whose names are used to make many ideas and definitions briefer.

A gate key is a start or resume key. Gates are used to address messages to domains.

A Segmode key is a segment, node, fetch or sense key. Segmode keys will serve to provide access to a segment defined by the designated node.

A memory key is a segmode key or a page key. Memory keys serve to define storage.

A primary key is any key that is not a gate key. Messages sent to a primary key are responded to directly by the kernel.