Smarts woos SA prospects

Smarts, the New York infrastructure management company represented in SA by Network Support Services, is touring the country to build on its relationship with Sasol and lure customers of HP OpenView, IBM Tivoli, CA UniCenter TNG and MicroMuse.

This market has recently seen solutions like Smarts' InCharge and Entuity's smaller Eye of the Storm making a play. Founded in 1993, Smarts focuses on automated faultfinding technology in complex enterprise environments.

Large companies have multi-vendor physical networks and computing infrastructure, applications, databases and business customers. "The inter-dependence of all these means infrastructure errors will propagate in complex ways that need to be managed and analysed in an integrated 'cross-domain' method, to provide true root cause analysis and business impact analysis," says Phil Andrews, Smarts VP for Europe, Middle East and Africa.

Smarts' core technology is the InCharge suite of solutions that deliver business assurance. It has been shipping directly and via partners for five years and has 130 customers, as well as over 3 000 customers through OEM partners like Cisco. Smarts is used in AT&T, Colt, Cable & Wireless, Microsoft and CSC, among others.

First Albany Research puts the cost of downtime for financial services organisations at R820 000 per minute. But downtime doesn't only mean lost business, it also includes loss of reputation, permanent customer desertion and more, meaning systems must be kept running.

"Finding the fault is the most difficult part," says Andrews. "InCharge provides one management dashboard with all company resources, their relationships and operating status. If you use a variety of management systems, each operating in a silo, to monitor various aspects of the business, this complicates matters.

"IT managers can receive thousands of alarms per day, some false," he adds. "InCharge can reduce alarms to genuine faults by as much as 95%. It takes in hundreds of thousands of events and diagnoses to the root cause across the network and infrastructure, and models the business impact."

The solution manages from the element level to the service level. InCharge's analysis layer is based on the common information model (CIM) standard, which "ties" all infrastructure modelling together. It is used to "predict the impact of faults and the behaviour of infrastructure when going faulty".

Infrastructure management systems operate at different levels of abstraction. At the basic level, there are networked elements (routers, switches, etc). Element management systems typically manage these, and are connected to monitoring solutions such as HP OpenView NNM and Tivoli Netview. In general, Andrews claims, these solutions give a view of devices impacted by outages, but little real actionable information to identify and isolate the root cause without significant programming at the time of the failure, leading to delays in resolution.

Implementation of these systems and the need to manage across multiple domains have necessitated further deployment of "manager of manager systems", such as MicroMuse NetCool, which makes it necessary to do a lot of rule programming at the network layer, for actionable information. This approach is complicated when the infrastructure or the interconnections change, as rules must be revalidated or changed.

To deliver a business view, the system has to interpret events from all silos. Smarts' InCharge CIM automation models behaviour automatically, eliminating writing of rules.

InCharge automatically discovers network and infrastructure elements and interconnections, models every possible fault and computes the symptoms, taking into consideration inter-relationships. These "fault signatures" are stored in the "codebook" - a matrix of each fault and the related signature. This process takes a matter of minutes and is the equivalent of man-years of "rule-coding". Any change in the infrastructure is re-learnt and the "codebook" is recalculated in minutes.

"Other management systems entail the user or IT manager setting up his own rules and keeping them up to date," says Andrews.

InCharge may cost between $120 000 and $2 million, depending on the size of the network. Larger networks have better use for it, given their complexity. Return on investment can take between three and nine months, says the company.

Sasol Polymers implemented the technology six months ago, bringing service level agreements in-house, as its main concern was plant uptime.