Oidview

Network monitoring, SNMP and network management guides from Oidview

What an SNMP OID Is and How to Read One

Written by

Published

Filed under

SNMP OID

Staring at a string like 1.3.6.1.2.1.1.1 and wondering what it means? An SNMP OID is a dotted number that gives one managed value on a device, such as its description, a unique place in a global tree.

Item Value
Internet subtree prefix 1.3.6.1
System description object (sysDescr) 1.3.6.1.2.1.1.1
Instance of a non-table object Name plus a final sub-identifier of 0
Maximum sub-identifiers per OID 128
Maximum value of one sub-identifier 4294967295
sysDescr text length limit 0 to 255 characters

Figures come from the SMIv2 specification and an OID repository entry. Standards and listings can change, so check the sources before you rely on them.

This page is part of our wider guide to SNMP, and it focuses on reading the numbers themselves.

What an OID actually identifies

SNMP does not ask a device for “the uptime” by name. It asks for an object identifier, a sequence of integers that names exactly one piece of management data.

The specification calls each number in the list a sub-identifier. Per the same document, there are at most 128 sub-identifiers in a value, and each one tops out at 4294967295.

The names you see in tools, like sysDescr, are friendly labels from a MIB file. On the wire, only the numbers matter.

The tree behind the numbers

OIDs are positions in a tree. Every number is a branch you take from the previous one, starting at a root and walking down to a leaf.

The SMIv2 specification defines the first steps, which nearly every SNMP OID shares:

  • 1 is iso, the first well-known root name.
  • 1.3 is org, the next branch under iso.
  • 1.3.6 is dod, the Department of Defense branch.
  • 1.3.6.1 is internet, where SNMP’s objects live.

Under internet, the mgmt branch (2) holds mib-2 (1), the standard set of objects most devices support. The private branch (4) holds enterprises (1), where vendors get their own space.

SNMP OID equipment
A white router with four upright antennas sits under blue and pink lighting.

Reading a real example, step by step

Take 1.3.6.1.2.1.1.1, the system description. A repository listing of this object shows it generated from the MIB-II definition, with the text “A textual description of the entity.” Read it one number at a time:

  1. 1.3.6.1 gets you to internet.
  2. .2 moves to mgmt.
  3. .1 moves to mib-2.
  4. .1 moves to the system group.
  5. .1 selects sysDescr within that group.

That listing gives its syntax as a display string of up to 255 characters, read-only. A device typically returns its model and software version there.

Why a trailing zero appears

You will often see the same OID ending in an extra number, as in 1.3.6.1.2.1.1.1.0. The definition names the object, but a query needs a specific instance.

For objects that are not in a table, the specification says an instance is identified by appending a sub-identifier of zero to the object’s name. That is why a GET for a single value usually ends in .0.

Table columns work differently. They append an index, such as an interface number, so one column OID yields one value per row.

Vendor OIDs and how to find the one you need

Standard objects sit under mib-2, but vendors extend the tree under enterprises. If a value is not in the standard set, its OID is probably under 1.3.6.1.4.1 followed by the vendor’s number.

To find an OID, load the vendor’s MIB file into a browser tool and search by name. The tool translates the name to numbers for you.

Which protocol version you use does not change the OID itself. For how requests and security differ, see SNMP versions, and for how queries reach the device, read about the SNMP port.

Numeric and named forms side by side

Every OID can be written two ways. The numeric form is a string of dotted integers, and the symbolic form joins names with dots, such as iso.org.dod.internet.

The two always point at the same object. Software converts between them using the MIB modules it has loaded, and only the numbers are sent inside a request.

This has a practical effect when troubleshooting. If a tool prints a raw number instead of a name, it simply has no MIB loaded for that branch, and nothing is wrong with the device.

  • Numeric output is always available and unambiguous.
  • Named output is easier to read but depends on the MIB files installed.
  • Mixed output, with a name for the known prefix and numbers after it, is common with vendor branches.

Common mistakes when reading OIDs

Most confusion comes from a few repeated habits. Knowing them saves time when a query returns nothing.

  1. Forgetting the instance. Asking for the object definition without its final zero often returns “no such object” or “no such instance” from a device.
  2. Treating numbers as sortable text. OIDs are compared number by number, so a longer sub-identifier is not “earlier” just because it starts with a smaller digit.
  3. Dropping a leading dot. Some tools accept a leading dot to mean “start at the root”, while others treat the string as relative to a default prefix.
  4. Assuming a vendor branch is universal. An enterprise OID only means something on that vendor’s equipment.

When in doubt, check the object’s definition in its MIB. It states the syntax, whether the value is read-only, and where it sits in the tree.

How OIDs are used in everyday monitoring

Monitoring systems poll a list of OIDs on a schedule. Each poll asks a device for the current value of an object and stores the answer for graphs and alerts.

A GET request names one or more instances and asks for their values. A walk starts at a branch and keeps asking for the next OID in order, which lets you list everything below that point without knowing each name in advance.

That ordering is why the tree structure matters. Because related objects share a prefix, a single walk of a group such as system returns all the descriptive details about a device in one pass.

It also explains why vendor OIDs are grouped under one number each. A tool can walk a vendor’s whole branch to discover custom sensors, fan speeds or licence counters that no standard MIB covers.

Frequently asked questions

Is an OID the same as a MIB?

Show answer

No. A MIB is a file that defines and names many objects, while an OID is the numeric address of a single one of them.

Why do OIDs start with 1.3.6.1?

Show answer

Those four numbers trace iso, org, dod and internet. The SMIv2 specification defines that path as the base for internet-standard management objects.

Do I type the name or the number?

Show answer

Most tools accept either if the right MIB is loaded. The numeric form always works, because that is what travels in SNMP messages.

What does the .0 at the end mean?

Show answer

It marks the single instance of a non-table object. Without it, you are naming the definition rather than a readable value.