Network Management Systems (NMS)

Integrate under a single umbrella all the management functions that are needed to efficiently administer a heterogenous datacommunications etwork.

Execute both routine functions (fault detection, problem tracking, data gathering) and sophisticated activities (continuously analyze network performace and report key findings).

Fault, Configuration, and Performance Management functions can be at low or high levels.

Little security or accounting tools

Major components of NMS

        Data gathering from all networkdevices or their proxies
        Graphical display that shows network status and allows logical
connections between various levels of hierarchies.
        Relational database  to store and retrieve all data
        Statistical tools for anaylysis of data
        Modular design to both expand and customize
        Trouble ticket system

NMS Integration Approaches

        Manager of Managers (MOM) -
        hierarchical
        links an integrated management system to existing element
management systems
        Open Management Platform (OMP)
        distributed
        uses open protocols to link management system directly to managed

elements

Criteria for selection

        Legacy systems
        Future of DCE in environment
        Client-server
        Cost and scale
        Importance of LANs vs. WANs
        Politics

Major Commercial NMS

        Enterprise-oriented
        Lan-oriented
        Market-leaders:  HP Openview   SunNet Manager   Netview/6000

HP Openview

        Chosen as the basis for OSF DME
        Uses SNMP for connectivity and fault management
        Hardware independent
        Distributed multiprotocol Management platform
        Uses Ingres database data

Network Node Manager Specifics

        User-defined threshholds for network devices with triggers
        Allows users and third party to create applications
        Weak application base
        MIB browser
        Dynamic mapping updates display and SNMP MIB
        Continuous Network Discovery

IBM Netview/6000

        Based on HP Openview, rewritten by IBM in object-oriented C++
        Uses filters and control interface for graphical display of SNMP events
        Excellent gui with zoom capacities

Customer Requirements

        Ease of use - windows, point and click
        Scalability - small networks to large
        Flexible functionality - global and local settings, time variances,
multiple user authorizations
        Customization tools - reports, engineering analyses
        Stringent performance criteria - reliability, response time, grace
under load
        Open interfaces to other NMSs - interoperability

NMS technical architecture I

        Interoperable Core systems
        multiple distributed monitors
        integrates with databases
                statistical and fault
                accounting
                asset management
                trouble tickets
NMS technical architecture II
        supports interfaces to other NMS
        forwarding alarms
        open API for applications
        Graphics System
NMS technical architecture III
        Network monitor
                hierarchical views
                alarm text windows
                context-sensitive help
                clickable diagnostic requests
        Performance reporter
                analytic tools
                chron jobs and real-time analyses
        Network builder
        intelligent rules
        resource discovery

Success factors for NMS

        The limits of network discovery
        Installation procedures
        Intelligent use of colors and annotation
        Scripting capabilities - customization and complexity
        Structure of information
        Demonstrable cost savings
        Reliability