How Arm cut datacenter onboarding time by 92% with the NetBox Labs platform
Share
0
0
0
9
9
0
0
0
2
2
0
0
0
%
%
less time to onboard a datacenter
0
0
0
~
~
0
0
0
2
2
0
0
0
0
0
0
h
h
0
0
0
r
r
0
0
0
s
s
to load a datacenter with hundreds of switches
0
0
0
3
3
0
0
0
+
+
new datacenters in six months
Customer Snapshot
A datacenter in about two hours
A discovery script loads every switch, interface, and IP address into NetBox in a single run, replacing days of manual entry.
Drift caught every few days
Scheduled rediscovery compares each device to its NetBox record and updates it, so changes surface early.
Provisioning from one entry
Plug in a device, give it an IP address, and discover it. Ansible playbooks take it from there using the NetBox record.
Arm's Global Networks team supports more than 2,500 network devices across campus networks and datacenters, for a company whose compute reaches 100 percent of the connected global population.
"Eliminating the administrative overhead, while gaining support and the enhanced feature set, it just made sense to use NetBox Cloud as our strategic tool for automation."
Jamie Lawrence Manager, Network Reliability Engineering, Arm
"Eliminating the administrative overhead, while gaining support and the enhanced feature set, it just made sense to use NetBox Cloud as our strategic tool for automation."
Jamie Lawrence
Manager, Network Reliability Engineering, Arm
The Challenge
Network data in three places, and none of them built to be the record
|
Hundreds
of switches in a new datacenter
| × |
Hundreds
of interfaces on each switch
| = |
Days
of manual entry before the network could be automated
|
Every new device was typed by hand intoAnsible YAMLServiceNow CMDBSpreadsheets
Jamie Lawrence’s Network Reliability Engineering team set out to replace all three with one automation foundation that could keep pace with Arm’s growth.
The Solution
One discovery run fills the record, and every provisioning job reads from it
 |
From a lab install to NetBox Cloud
Started in the lab
Another Arm team ran NetBox Community for lab environments. Jamie’s team saw it could handle the whole corporate network and adopted it as the automation backbone.
Outgrew self-hosting▼Running NetBox Community in Docker meant owning upgrades, hosting, and disaster recovery as the team relied on it more.
Moved to NetBox Cloud▼NetBox Cloud took the upgrades, hosting, and disaster recovery off the team’s plate and added dedicated enterprise support and an enhanced feature set.
|
|
Discovery that fills the record
See how NetBox Discovery works One script, every detail
A custom script connects to each device and pulls model and serial numbers, interfaces, and IP addresses into NetBox.
Hundreds of switches in about two hours▼A full datacenter now loads in a single run, work that used to take days of entry by hand.
New device types handled automatically▼When an unfamiliar model appears, discovery creates it in NetBox without anyone defining the device type first.
|  |
 |
Provisioning and drift checks from one record
Talk to our team about automation Ansible reads from NetBox
Playbooks query NetBox for device details and push standard SNMP and authentication configuration.
Rediscovery every few days▼Scheduled runs pull any change in device state back into NetBox, so drift shows up early.
No repo caretaker▼With discovery keeping records current, no one has to maintain the device repository by hand.
|
|
A platform that fits the workflow
Explore the NetBox Labs platform Custom fields and config contexts
The team shapes NetBox around its own workflows with custom fields, configuration contexts, custom buttons, and JSON.
Enterprise support, still customizable▼“It’s actually really simple things, but it does the simple things well. And that’s the key.” Joshua Jacobs
A direct line to product teams▼Arm asked for direct access to product teams from the start, and that feedback loop cemented the partnership.
|  |
"Using NetBox is much more efficient and it speeds up our engineering time. So if we're deploying a new device, all we really have to do is plug it in, give it an IP address and then we discover it into NetBox. Then we can run the rest of our provisioning scripts and all it needs is that entry in NetBox and we're done.
Joshua Jacobs
Automation Engineer, Arm
The Results
Days of manual entry became a two-hour run, and the network keeps growing
|
Automated discovery and provisioning
~2 hours for an entire datacenter with hundreds of switches, work that previously took days
|
92%
less time to onboard a datacenter, from days of manual entry to one run
30+ min
saved every day now that rediscovery keeps the inventory current, with no one assigned to maintain it
2,545 → ~3,000
devices today, and the projection for year-end as new datacenters and halls come online
|
Next for Arm’s Global Networks team
NetBox Discovery and NetBox AssuranceNative auditingCatalyst Center, Infoblox, ACICloud firewalls and routers
“Our OKRs are very AI automation oriented. And so NetBox is, and needs to be, that foundational pillar to help drive where we’re going with automation.”
| Jamie Lawrence Manager, Network Reliability Engineering, Arm |
Get the full Arm storyDownload the PDF
Start your automation journey
