Transcript Chapter14 - YSU Computer Science & Information Systems

Chapter 14
Managing E-Business and
Network Systems
Introduction
Managing e-business systems and their
underlying infrastructure is a critical
success factor for managers
In the e-business world, networks are
the key ingredient linking, systems,
processes, and people
Networks add value and complexity to
information infrastructures
System Operational
Disciplines
Managing Batch Systems
Batch systems processing involves receiving
and aggregating incoming transactions and
distributing the resulting output data
Regularly scheduled applications
Applications are commonly executed on centrally
located computer facilities
Accounts payable, inventory, ledger
Applications commonly use computer scheduling
because of interjob dependencies
Network Systems
Management
Networks form the basis for many well known
and emerging businesses
Amazon, eBay
Network technology enables and encourages
restructuring and re-engineering processes
Web technology adds new dimensions to a
firm’s systems and management processes
Network Management’s
Scope
Network management’s focus is broad and
diffuse
Managers are responsible for owned as well as
leased assets
Boundaries are also blurred between data and
voice applications
Network managers must provide seamless
support for customers while tracking and
solving problems across the enterprise as
well as those involving outside vendors
Managers’ Expectations of
Networks
Growth in networked applications demands
increased network management capability
Users expect networks to be capable,
reliable, and cost effective
Unfulfilled expectations are an important
source of difficulty for IT managers; networks
can be a prime source of failure
Network Management
Disciplines
Disciplined techniques are critical for network
management success
SLAs incorporate customer expectations of
reliability, responsiveness, and availability
Managers must focus on problem, change, and
recovery management
Performance planning and analysis, capacity
planning, and configuration management are also
important
The Disciplines of Network
Management
Network Service Levels
Users view networks as a unified entity, so
the SLAs must treat applications, computer
and network hardware, network links, and
user workstations as an integrated whole
SLAs must include measures of availability,
service quantities, and reliability
They must include some measures of workload
Configuration Management
Configuration management includes a database
containing an accurate record of the network’s
physical and logical connections and
configurations
Configuration management’s scope includes
Physical Connectivity
Logical Topology
Bandwidth
Equipment (inventory and specifications)
User Information
Vendor Data
Fault and Change
Management
Fault management is similar in many ways to
problem management
Networks can be created with monitoring to
automatically flag failures and attempt to
reconfigure
Faults can arise from hardware, software, or
configuration failures
Troubleshooting these failures requires the skills
of network engineers and technicians and access
to configuration databases
Recovery Management
Network managers must plan to recover from
local disasters as well as disasters affecting
vendors
Redundancy in name (using different carriers for
reliability) may not be redundancy in fact (the
carriers use the same physical ROW)
Managers must remain aware of this problem and
explicitly address it when initiating SLAs with
telecom vendors
Network Management
Systems
Automated tools that help manage and
operate networks
Gather statistics from routers and switches
SNMP traps
Exist as hardware embedded in the network to
create diagnostic logs
Monitor network usage and performance
Vendors offer integrated monitoring and
management packages
Performance Management
Techniques for defining, planning, measuring,
analyzing, reporting, and improving on
infrastructure performance
Defining performance
Performance planning
Measuring performance
Analyzing measurements
Reporting results
System tuning
Defining Performance
System performance is the volume of work
accomplished per unit of time
CPU throughput, network transmission bandwidth,
number of transactions posted
In e-business applications, system response time
is a critical performance measure for end user
satisfaction
With rapidly improving hardware performance,
managers are shifting from increasing HW
efficiency to improving end user satisfaction
Performance Planning
Establishes objectives for human/ computer
system throughput
Workload characterization is the cornerstone of all
performance and capacity programs
System performance and associated factors must
be well understood prior to system capacity
increases
System tuning (optimization) can also yield
performance increases without capacity expansion
Measuring and Analyzing
Performance
Measuring response time and system
throughput under a variety of workloads is
critical
Transaction service time
Transaction rates
Average response time
These measurements are used to judge
delivery of SLAs, capacity trends, and tuning
results
Network Performance
Assessment
Network managers must monitor
performance to exceed SLAs
As usage and loading changes, network
bottlenecks occur
Without some system of monitoring, resolution of
these problems is delayed with user satisfaction
declining
Network availability is a calculation derived from
MTBF and MTTR
Availability = (MTBF)/(MTBF + MTTR) X 100
System Tuning
System tuning or optimization can be used to
create performance increases without capital
expense
Risks of performance tuning include the risk of
change as discussed earlier
Sometimes limiting access to a resource, while
decreasing capacity, improves throughput
Tuning of complex systems can be very time
consuming and tedious
Capacity Management
Process by which IT managers plan and
control the quantity of system resources
needed to satisfy user needs
The goal is to match available system
resources with those needed to meet service
levels
Must also anticipate future needs and plan for
increased usage
Must also identify obsolete or underutilized
hardware and services
Capacity Analysis
Managers must perform a detailed analysis of
current system resource requirements
Acts as a benchmark for proposed changes
Needs to identify daily workload peaks as well as
peak loading for weekly and monthly timeframes
Capacity assessment and monitoring must be a
continuous process in rapidly growing ebusinesses
Capacity Planning
Managers must anticipate future need
and plan accordingly
Techniques range from simple to complex
Managers must pick the correct metrics to
follow, as technology changes so do needs
Planning must be grounded in fact and
logic, not hunches and guesswork
Additional Planning Factors
Changes in the organization’s strategic
directions that might modify or increase IT
services
2. Business volume changes (either increases
or decreases)
3. Organizational changes (always a potential
impact on IT resources)
4. Changes in the number of people using IT
services
1.
Additional Planning Factors
Changing financial conditions within the firm
or industry
6. Changes in service-level agreements or
service-level objectives that might have a
bearing on system performance
requirements
7. Portfolio management actions that might
impact system throughput, such as the
addition of new applications or
enhancements to current applications
5.
Additional Planning Factors
Testing new applications or making
modifications to current applications that
require additional system resources
9. Application schedule changes initiated by
operations or user managers
10. Schedule alterations for system backup and
vital records processing
11. System outage data and job rerun times
from the problem management system
8.
Linking Plans to Service
Levels
Periodically the performance and capacity
management processes must be reviewed to
assess their effectiveness
Did forecasts agree with actual demand?
Did capacity needs match capacity resources?
Were service levels met?
Were budgets adequate?
Are customers satisfied?
Management Information
Reporting
Reporting is an essential role – it creates
transparency to IT operations
The intent of reporting is to
Improve operations
Promote organizational learning
Engage customers in dialog that generates results
Communication increases trust and confidence
between participants
Reports are essential not only to providers but
also customers
The Network Manager
Network managers need to be skilled
generalists
Their duties transcend organizational, political,
cultural, and geographic boundaries
Technology adoption enables and mandates
structural changes
They must be capable tacticians, understanding
where they need to take the firm and how to get
there
Summary
Centralized batch systems and network
applications depend critically on network
performance and computer resources
There exists a strong link between system
performance and system capacity
Quality IT business processes must be an
overriding consideration in the IT organization