Why Are Network Changes So Risky?

Industry statistics reveal that 3 out of 4 network outages are caused by human changes to the network rather than hardware failure or circuit outages.  In other words, a perfect monitoring system will at best only catch 25% of network issues.  So what can we do?

Solution 1: Invest in engineering processes – Require more design reviews before changes, more testing after changes and carefully control when network changes can be made and who can make them.

Solution 2:  Invest in configuration management tools – Catch unscheduled changes and prevent similar situations from happening again.

Although most large enterprises have both solutions in place today,  the problem is still largely unsolved.

Time for New Thinking

There are countless ways network changes can go wrong.  The sheer complexity of the various tasks involved demand an Integrated Solution that can combine human wisdom, computational intelligence and automation tools together.  This novel idea is the inspiration behind NetBrain’s Map-Driven Network Design environment.

There are three key-components in NetBrain’s integrated solution:

• Enable collaboration with a map-driven design environment

NetBrain automates the creation of design documentation, from Visio network diagrams to comprehensive design documents in Word.  As a result, it can greatly increase the efficiency of the human review process, which is by far the most effective way to increase design quality and reduce outages.

• Simulate network design before implementation

A large portion of network complexity is sourced from dynamic interaction between network devices.  Things like dynamic routing and traffic filtering are difficult to replicate in a lab or sort out in the human brain.  Network simulation leverages computational intelligence to help people predict what-if scenarios.  With the ability to define and validate an optimum network design, you no longer need to anxiously await uncertain network change impacts.

• Automatically verify impacts after implementation

Networks are constantly changing… Every new modification will induce supplementary changes due to the ripple effect inherent with networks.  Automated verification will discover changes in routing, configuration, topology and even traffic path automatically right after the network changes.  Leave no stone unturned, so you can be sure that your network is running smoothly.

NetBrain Workflow for Integrated Network Design and Implementation

How It Works

Inside NetBrain, everything starts and ends with a SmartMap. SmartMaps are the 3rd generation network diagramming technology that allows engineers to create dynamic network diagrams on-the-fly, by leveraging an underlying data model containing Smart Objects.

To understand the map-driven network engineering environment, take a look at the following examples:

Objective: Re-design the routing between two POPs (Point of Presence) so that traffic between them goes through the core network, instead of the client network.

SmartMap created instantly for the proposed design

Map-Driven Network Engineering Workflow

Design	Plan	The planning process inside NetBrain for network changes may include: Discovery of the portion of network to be changed, such as a data center Creation of dynamic network diagrams about proposed network changes on-the-fly Building new smart objects(if new devices are added), or modification of network configuration files (only if configuration such as ACL is to be changed) For this example, the following configuration change for the NY_POP router is proposed:
	Simulate	Utilizing the simulation engine is a powerful way to predict what will happen without resorting to a costly labs. The software will take into account all network configurations such as dynamic routing, ACL, and route maps to compute the routing table for each device. Issues such as over-looked routing propagation can be caught before touching the live network. The proposed configuration changes above overlooked the detail of distribution list. The simulation below shows that the traffic flow will go to a firewall and create a permanent routing loop.
Implementation	Benchmark Before	After the design proposal is fully evaluated, changes are ready to be deployed.  NetBrain provides the one-click benchmark capability to document network status before network change for any part of a network, right from a SmartMap.  The data you can benchmark includes: Configuration files Routing tables Show-command results Topology
	Execution	Making live configuration changes with NetBrain, or other 3rd party tools.  NetBrain provides automated login capabilities to any network device it manages.
	Benchmark After	Right after the changes are executed, use SmartMap to take a snapshot of the live network again to prepare for the next step of analyzing the impacts.
	Automated Verification	What exactly has been changed in the live network? This question will be answered thoroughly in this step. The comparative analysis of two benchmarks (before and after change) will report: Configuration changes occurred Routing table changes occurred Traffic path changes Any show-command result changes occurred Topology changes occurred
Documentation	Document the entire change process	With NetBrain, it takes minutes rather than days to document a complex change. In fact, once the design or implementation is complete, all data is automatically saved along with the SmartMap. If you need traditional Visio diagrams or Word formatted network documents, they are only one-click away. See a sample design document here.

See It in Action

• Data Center Migration
• Integration of Two Networks After Merge and Acquisition
• Management of a Regular Network Upgrade

What’s Next...?

• Success Story