Not long ago, a company
asked me to review its IT documentation as it prepared for an audit.
Systems, hardware,
and software documentation were in order, but the network was a hodgepodge of
documents that failed to connect the dots. Instead, the story that network
documents revealed about the network was that many changes had been made
quickly to accommodate another organization that had been acquired and to add a
multitude of new users in different locales.
Often, companies
find themselves in this situation. They undergo change rapidly, and it’s
difficult to keep network documentation up to date. More troubling is the
inability to develop a clear-cut network roadmap that addresses future network
needs, expansion, and security.
How do you develop
an effective network roadmap while keeping pace with workloads? Proceed step by
step. Building a network roadmap is an exciting strategic exercise, but first
the current baseline network should be assessed and understood.
There are numerous
steps in this process. Undertaking these steps while concurrently managing
daily network health and operations can be daunting.
Because of this,
it’s useful to break down network roadmap building into a series of systematic
steps, and to tackle each step individually.
Step 1: Revisit
performance and workload goals
What network
performance goals must be met each day for the business? Where are they being
met, and where are they failing? Are your network goals variegated? For
instance, an organization might have an internal network for employees and an
external telemedicine network for doctors and patients. Each is likely to have
different performance, throughput, and quality of service (QoS) goals.
By reviewing
performance metrics and trends, you can see where the network’s strengths and
weaknesses are as you build the roadmap.
Step 2: Automate
documentation
The future of the
network is automation.
This should start
with what the Achilles’ heel of most networks is likely to be: documentation.
Most network
documentation is still largely manually maintained and out of date. Network
administrators often find themselves working like CAD engineers, developing
network schematics from scratch — if their daily work allows them time to do
this.
One solution is
automated network documentation. There are tools on the market that will scan
your network and produce their own network schematics, which you can
tweak or add notes to as needed.
Step 3: Chart
your network monitoring direction
Everyone wants to
go to 5G, boost Wi-Fi, support on-demand video, audio, data payloads and a
plethora of devices and platforms, but how do you monitor networks at this
level of sophistication?
The path suggested
is network automation, which moves from manual network monitoring to automated
monitoring to network observability.
Today, companies
use a mix of manual and automated monitoring. Network professionals check dashboards
and drill down into trouble areas. There is also software monitoring and
automation software that scans network entry points for vulnerabilities and
security breaches, provides traceability, and performs automated functions like
updating software to the latest versions across all devices on the network.
Some of this software monitors IoT (Internet of Things) edge technologies,
issuing alerts when network anomalies or abnormalities are detected.
Unfortunately,
when network issues are found, network professionals must drill down into them,
determining root causes so they can be solved. This is where the next
generation of network monitoring — observability — comes in.
Step 4: Define
your network automation path
If automation is
the path of network evolution, what most companies want to do is to automate
more elements of monitoring and progress into the stage of network
observability, where artificial intelligence (AI) can tell you not only that
something is wrong, but why it is wrong. Using observability software,
which employs machine learning (ML) to learn the dynamics of your network
infrastructure so it has a context for issue troubleshooting, can speed times
to problem resolution because the observability software not only issues alerts
— but it also tells you why the alerts might be occurring based on what
it knows and what it has observed. This cuts down manual root cause analysis
time.
Observability uses
network logs to determine when specific events connected to an issue occurred,
who or what generated the issue, etc. It tracks metrics such as how much memory
or bandwidth was used by a certain network request — and it traces events as
they move from one network node to the next.
AI-based
observability can winnow through a slew of network alerts, isolating only the
ones that matter, because it understands the network’s operational
infrastructure. It can suggest root causes for what is wrong. This is what
speeds time to resolution for network professionals and is the future of
network automation.
Step 5: Decide
what’s next
If most companies
recognize that observability and full automation of the network is their
network monitoring future, the key to building the network roadmap will be
filling in all the points between where the network is now and the endpoint of
what could be total network observability.
It also means that
the most likely steps to be taken are gradual progressions in automation and
phase-ins of what will likely be substantial funding needed for network
automation tools.
In all cases, it
is likely that observability and total automation will be tested incrementally
to build trust in the technology, and to define the intervention points for IT
staff in newly automated network processes.
Related
Content:
Is
Networking the New Killer App?
University CIO Modernizes
Networking Infrastructure
The Deep Skills Every Network Manager Should Possess
Source link
