Table of Contents
Advertisement
Design for High Availability
As enterprises accelerate their migration from traditional circuit-switched telephony services to
IP Telephony solutions, a major consideration is their ongoing concern as to the potential
reliability of the proposed IP voice services versus that of their current infrastructure. Indeed, in
many call center environments, the potential cost of downtime is often greater than the implied
benefits of migrating to IP Telephony. This chapter outlines an analytical approach to projecting
the availability of an enterprise's IP Telephony solution by examining the characteristics of the
critical components of the telephony system and the traffic handled by different subsystems.
The availability projection will assist with designing a configuration that will meet the enterprise's
reliability expectations.
Availability of a switching system is traditionally defined as the fraction of time that the system is
operational. This metric is generally calculated from the end-user's perspective and does not
necessarily reflect the frequency of individual component failures or the maintenance required.
Despite the fact that most people can empirically assess whether a device or service is
operational, the telecommunications industry entreats us to examine these availability issues
with a high degree of rigor, structure and methodological comparison. This section walks
through several typical configurations to show where available redundancy options can be
implemented to improve availability. As the first step, the availability of the critical components
that comprise the enterprise's IP Telephony system, such as the supporting hardware, software
and the underlying data network infrastructure must be analyzed. Then, the full system
availability is calculated based on the estimated traffic generated at each site and by each
subsystem.
In addition to the aforementioned parameters, there are other major contributors to failure of a
telephony system. These include outages due to user error, PSTN (Public Switch Telephony
Network) failure, and external power outages. The downtime experienced due to these types of
outages is not considered when projecting the system availability because they have an equal
likelihood of impacting all solutions. Outages as the result of "user error" and improper network
design can be significantly reduced with proper design and good networking practices.
Sometimes recommended components for mitigating power outages, including redundant
power supplies, backup generators, and Uninterruptible Power Sources (UPS) are simply
overlooked or not acted upon due to their cost. The Public Switched Telephony Network (PSTN)
in the United States is expected to meet 99.997% availability per year and, as such, often
serves as the de facto standard of availability for business applications involving voice
communication.
impact on its telephony system's availability. For example, for a contact center with 85% of its
traffic being outbound calls to PSTN trunks, the expected full system availability can never
actually be greater than that provided by its supporting PSTN.
3 D. Richard Kuhn, Sources of Failure in The Public Switched Telephone Network, IEEE Computer Magazine, Vol.
30, April 1997, pp. 31-36. "For several decades, AT&T has expected its switches to experience not more than
two hours of failure in 40 years, a failure rate of 5.7x10
3
Relative to the enterprise's type of business, this number can have varying
Design for High Availability
-6
."
Issue 3.4.1 June 2005
255
Advertisement
Table of Contents
