November 2003
A New Era For Testing And Monitoring Tools
BY WAYNE R. FULLER
IT departments are facing a dilemma. These organizations, once relegated
to the basement operations of organizations, are now seen as vital to the
well being of every level of business. These departments provide the
critical backbone through which all communications with customers,
suppliers, employees, and management rides on -- the lifeline of any
business. They are responsible for everything from applications availability
to network performance. But, with this elevated visibility in the corporate
structure, comes exposure to serious cost analysis of IT department
spending.
Fortunately, enterprise IT organizations have more options than ever before
to reduce the costs of their communications budgets. Many new technologies
and applications present opportunities for the IT department to reduce these
budgets, while at the same time improving the efficiency of their
operations. Converging voice, video, and data over an IP network is one
option. Voice over IP has the potential to save organizations millions of
dollars in long-distance expenses.
The goal then for the IT department is about more than just evaluating and
deploying new technologies, it�s about ensuring that the technologies they
deploy as well as the current infrastructure, are optimized, reliable, and
cost effective. Before VoIP can be an effective and cost-saving business
tool, it needs to offer the same reliability of landline phone.
The dilemma for IT organizations is this: How do they cost effectively
deliver services -- whether they be traditional or next-generation -- to
their �customers� without sacrificing performance?
Historically the answer was simple, IT organizations would utilize a
plethora of testing and monitoring point solutions to ensure that not only
guaranteed service-level agreements (SLAs) from their carrier were being
met, but that they were also delivering the appropriate level of �service�
to their �consumers,� whether they be external customers, suppliers, or
internal business units. Occasionally, with or without testing and
monitoring tools, an enterprise IT organization may have opted for upgrading
network bandwidth as a solution for guaranteeing optimal performance of the
network.
Today, however, the answer is not so simple. The scrutiny being placed on
all IT spending, even for network and performance management solutions,
becomes harder and harder to justify. Furthermore, IT departments have to
sort through the myriad network and performance management point solutions
on the market to find ones that satisfy their specific service requirements.
Therefore, the cost benefit of using converged networks for voice, video,
and data traffic will only be met if an enterprise has the right solution
that fits its budget to engineer the network so that business critical
applications and delay-sensitive traffic are not adversely affected by low
priority applications. Moreover, enterprise customers need true end-to-end
performance management, not simply a solution that validates SLAs between a
service provider�s IP demarcation points.
For their part, enterprise customers will have greater confidence in
migrating to this converged world if they could get the kind of QoS
guarantees for IP services that they have come to expect from traditional
data services, like frame relay. Since IP connections do not have an
assigned amount of bandwidth as does a PVC that comes with a CIR (committed
information rate), enterprises would have to be certain that not only could
their mission-critical applications run smoothly over an IP infrastructure
but the low-delay and jitter requirements of their voice and video traffic
-- both of which can be more economically supported over a meshed IP network
-- can be met through an IP network.
NEW REQUIREMENTS
Today enterprise IT organizations are requiring alternatives to current
performance management deployment strategies. They need a highly
cost-effective, scalable, and flexible approach to network service testing
and monitoring that is targeted to their specific needs and service
requirements. Performance tools need to support different classes of
services associated with different types of communications, independent of
whether the traffic crosses private networks or the Internet.
Problems need to be quickly isolated; therefore solutions must have
real-time monitoring capabilities to the access line, access channel,
circuit, or the application. Testing and monitoring tools must also be able
to perform accurate bandwidth sizing to account for burstiness, as well as
report on SLAs for end-to-end connectivity and Class of Service (CoS).
Perhaps even importantly, an �open� system is required, which integrates
with other key aspects of managing a network and overall business processes.
Enterprises evaluating network and performance management solutions should
expect a tool that does more than merely use SNMP to communicate with WAN
access devices at various sites to collect basic status data. For example, a
performance tool capable of inserting traffic in the data stream can measure
network roundtrip delay. If customers are to put all of their traffic on an
IP network, then they need to have the kind of information that tells them
whether traffic congestion occurs at a specific time and what causes the
network to slow down. They need to know, for example, if network congestion
is due to database backups or is Web surfing by employees. Customers will be
able to tell what applications are slowing down the network if they have a
performance management system and intelligent WAN access devices that allows
them visibility into the network as well as into the IP traffic.
DEMANDING OPTIONS
IT organizations should be able to buy testing and monitoring tools in a
manner optimal for them. Enterprises want to purchase the features they
need, when and how they want them. Flexibility in pricing is key. Customers
should be able to get such a solution through a pricing plan that suits
their budget and they should not have to pay in advance for features that
they might want only at some later date. For instance, if customers decide
to put data traffic on the IP network before using it as transport for voice
traffic, they should not have to pay up-front voice monitoring capabilities
until they need such a feature but rather have a migration path to these
enhanced services as their requirements change.
By enabling initial entry at a modest price point and providing
functionality in an easy to deploy, modular approach, any enterprise can
utilize testing and monitoring tools based on their specific service
requirements. Solutions should be offered in such a way that enterprises
purchase based on immediate requirements, but can easily purchase enhanced
capability, as they need it.
An open performance management solution will allow customers to fully
leverage the capabilities of the product while retaining the option of using
multi-vendor WAN access devices. Users should not be locked into having to
install a particular vendor�s WAN access devices to monitor the network;
rather, they should have the choice of exploiting the features of the
performance management system using leading vendors� WAN access devices. At
the very least, such a solution will offer monitoring of any SNMP enabled
device and it will also allow for vertical integration with the
best-of-breed network management software, WAN analysis products, reporting
tools, and the like.
The importance of flexible options tailored to various customer needs should
not be overlooked because not every enterprise customer will want to, or is
able to, purchase an �all or none� performance management tool. In today�s
networked world, it�s all about choice. Giving the customer the option to
rent or lease a solution for period of time, either for some or for all of
their sites is an entirely new way of looking at testing and monitoring
solutions.
For instance, a customer might initially be satisfied with a real-time view
of the network but later decide to add on a feature set that could offer a
longer-term snapshot of network trending and performance statistics allowing
for further network optimization. Customers should be allowed to decide what
features of a performance management they want to use, equally importantly
for what period of time they want to use them, and with what third-party
management tools they want to integrate their solution. Customers running
data traffic on the network should have the option of purchasing a VoIP
monitoring tool only at such a time as when such a need arises, or, to use
another example, at their discretion turning on a security management
feature that can detect DoS (denial of service) attacks.
Migrating to new applications and technologies can be good for the
enterprise network and good for business. An architecture that offers a
cost-effective, flexible approach to performance management is the key to
realizing the inherent cost savings and efficiencies of new services without
sacrificing performance. No longer will enterprise IT managers be faced with
the dilemma of choosing between cost savings and innovation or performance.
With performance management solutions that allow entry at a modest price and
provide additional capability and functionality in an easy-to-deploy modular
approach, any enterprise can utilize powerful performance management data --
when they want and how they want -- based on their specific service
requirements.
Wayne R. Fuller is the executive vice president of operations at
Visual Networks, an IP performance
management vendor in Rockville, Maryland. Mr. Fuller leads the engineering,
technical services, information technology, and manufacturing groups at
Visual Networks.
[ Return
To The November 2003 Table Of Contents ]
|