Solution Overview
Key Benets
Telecommunications
One of the world’s largest telecom providers needed to replace an aging scale-out system with a
modern storage infrastructure to handle the immense volume and velocity of machine log data,
which is accessed by various analytics tools including Splunk, for its global network. QF2, a modern,
highly scalable file storage system, delivered the necessary capacity, reliability and performance,
with improved data visibility and customer support, all for dramatically lower cost.
Finding the Right Solution to
Replace Critical Infrastructure
The world’s largest telecommunications
providers make billions of daily endpoint
connections, and each connection generates
log data that must be ingested, stored and
processed to identify events or anomalies.
These logs represent terabytes of machine
data each day, multiple petabytes in total.
Storing that immense volume and velocity
of data requires incredible capacity and
performance, as well as 100% availability,
given the need to analyze data in real-time on a
continuous basis using analytics tools. Constant
uptime takes both reliable equipment and
extreme customer service.
That was the challenge facing a top
U.S.-based carrier.
“When you’re ingesting terabytes of data each
day from more than 60 billion incoming events,
you need storage capacity and performance
that’s very, very scalable,” notes a high level
executive. “You can’t afford a misfire with the
production system.”
The relevant data includes various types of
machine generated events such as server
syslogs and numerous router telemetry events
from many networks. The team stored that
machine data on an Isilon system that had been
in continuous use for five years. In fact, it hadn’t
even been offline in the last three years, for fear
the aging system might not come back up.
Knowing it needed to be replaced, the initial
thought was a substantial upgrade. However,
after the EMC acquisition of Isilon, the Isilon
product roadmap, service and support weren’t
meeting expectations and were deemed a high
risk to further consider as an option. With their
current vendor out of the running, the team
started looking elsewhere for a new system.
The company decided to evaluate object
storage to address scalability and leverage the
open programmability provided via an API. They
quickly learned that sacrificing the standard
file protocols, such as NFS, that the business
had come to rely on was a daunting task and
compromised daily operations. Ultimately, the
team needed to find a solution with the scalability
and API programmability of object storage
without sacrificing the proven reliability of a
standards based file storage system.
Qumulo File Fabric (QF2) Helps Leading
Telecommunications Provider Process Logs
from Billions of Connections Per Day
• 40 Qumulo QC208 hybrid
storage appliances
• 8 Qumulo QC24 hybrid
storage appliances
�� NFS and REST protocols
• Qumulo Care enterprise support
• Achieves the extreme
performance needed to handle
daily loads of billions of packets
and tens of thousands of IOPS
• Delivers usage statistics enabling
transition from closed vault to
open global intra-department
storage platform
• Scales throughput and capacity
linearly to support theoretically
limitless terabytes/day growth
with additional nodes
• Provides real-time data analytics
independent of scale for
immediate insight on massive
capacity and performance trends
• Ensures uninterrupted operation
through continuous Qumulo Care
monitoring and proactive support
• Adapts to demanding requirements
of even the heaviest production
environments through agile and
continual software evolution
“Right now, QF2 is the closest
thing to an Apple unboxing,
setup, and support experience
in the storage world.”
— High-Level Executive
Top U.S.-Based Carrier
Global Telecommunications Carrier Case Study
Qumulo, Inc. 1501 4th Ave. Suite 1600 Seattle WA 98101 1-855-478-6856 | info@qumulo.com | qumulo.com
The team was now left with an aging unclear
product future of Isilon and an object solution
that didn’t meet their current operational
requirements. Fortunately, they also had an ace
in their back pocket. The company had spent
the last several years watching the progress of
a new modern storage alternative as it made its
way to the market: QF2.
A New Gold Standard for Storage
“Right now, Qumulo’s the closest thing to an
Apple unboxing, setup, and support experience
in the storage world,” explains a high level
executive and senior engineers on the team.
���We loved the innovation and energy of Isilon’s
early days, and that’s what we saw in Qumulo. In
fact given all the attention and promise, our only
question was ‘could they deliver’? Fortunately
we found the answer was a resounding yes.”
Qumulo is the leader in highly scalable file
storage. It delivers real-time visibility into
data usage and storage across fast, flexible
commodity hardware. This combination
offered the storage performance and scale the
company required, along with the flexible API
based programmability the object system had
promised. As importantly, QF2’s hybrid solution
was significantly less expensive than either of
the other systems.
The carrier initially selected Qumulo’s QC208
hybrid storage appliances for its primary “gold
master data” cluster, deploying 20 nodes for
more than 4PB of raw storage. These systems
receive all incoming data, which is centralized
and made available to the various analytics
tools including Splunk. Centralizing all of the
data on QF2 gives the carrier the flexibility
to utilize a variety of tools for analysis. It also
deployed eight QC24 nodes — two four node
clusters of which serve as a QA test bed for
changes to the primary system.
Ease of setup was very important to the team,
as installation and cable runs are all handled
in-house. That wasn’t a problem with the QF2
system. “Turns out we had the initial four node
cluster set up before the notification even went
out to Qumulo confirming it was on the loading
dock,” notes a senior engineer on the team.
Given the degrading service experience
for its previous Isilon system after the EMC
acquisition, Qumulo was also under the gun to
show how nimble it could be – something that’s
much easier with agile two-week development
and release cycles for OS updates. “During the
sales process, we realized we needed support
for Ethernet jumbo frames. Our Qumulo sales
rep spoke with development, and the company
had it for us in the next OS update,” he says.
“That was amazing.”
Centralizing Machine Data on QF2
The carrier migrated all its data over a 30-day
period to the new QF2 production cluster, which
currently hosts almost 2PB of machine data and
is growing at a rate of more than
2.4TB per day with a two year retention period.
While QF2’s scalability is key to handling that
capacity load, it’s also critical for meeting
the extreme performance requirements —
particularly for data reads. A steady stream of
log data means high, but manageable, data-
write requirements. The system frequently gets
pounded for reads, as multiple simultaneous
processes and various analytical tools analyze
the data looking for anomalies or changes,
while applying metadata tags and other details.
QF2’s real-time data visibility is also increasingly
important for the carrier, enabling the team to
monitor and manage usage, and see capacity
and performance trends. It also supports the
team’s objective to move from the previous
‘vault’ storage mentality to an open platform that
can be used by other groups within the carrier.
QF2 gives real-time insight on when files get
committed, which groups are using what space,
etc., all supporting clear communication and
easy internal charge-backs.
The clean, intuitive web interface helps with
that usability, as does QF2’s REST-based
API. “Honestly, the management process is
painless,” notes the senior engineer. “I couldn’t
ask for things to be any easier.”
Qumulo has also been instrumental in ensuring
reliability and adaptability for this massive
system. Qumulo’s staff and executives regularly
meet with the team to discuss current and
future needs. And the Qumulo Care support
organization constantly monitors the carrier’s
cluster for potential issues.
The team is so pleased with their QF2
cluster that they have already expanded the
deployment with purchase of another 20
nodes (4PB) to mirror the production cluster
for disaster recovery, bringing the total current
installed base to 48 nodes.
Given the success of this deployment, the
organization overall now intends standardizing
on QF2 for its machine data storage worldwide.
“I’m reassured that as we grow, QF2 is right
there with us every step of the way,” he
concludes.
Global Telecommunications Carrier Case Study
Q147 0817
“When you’re ingesting
terabytes of data each day
from more than 60 billion
incoming events, you
need storage capacity and
performance that’s very,
very scalable. You can’t
afford a misfire with the
production system.”
— High-Level Executive
Top U.S.-Based Carrier
Qumulo, Inc. 1501 4th Ave. Suite 1600 Seattle WA 98101 1-855-478-6856 | info@qumulo.com | qumulo.com