 |
|
|
Redundancy is built into our entire network architecture
starting with redundant OC-3 fiber connections for voice and data provided by
Verizon and TelCove. We use two (2) long distance carriers to load balance
inbound and outbound voice traffic. A telephony server farm is used to handle
inbound calls. Service interruption when a telephony server fails is limited to
the calls that the server was processing while it failed. A fully redundant NAS
is used for storage of voicemail/fax messages and greetings. A SQL Server
Active/Passive cluster is set up for database redundancy. If the Active SQL
Server fails, traffic is automatically redirected to the Passive SQL Server.
Redundant systems are in place at all levels of our network in order to assure
the highest level of availability. Based on the second half of 2005, our
availability rate is 99.94%.
|
|
|
For
the Technology Geek in all of us, here are some specifics on the technology
behind our service, the Infrastructure, Security and Maintenance.
If you don’t have a strong desire to understand the entire behind the scenes
aspects of TeleCenter than just know this, we’ve got you covered!
Here
are a few pictures of our secure Data Center.
|
|
|
The COA Network platform was designed to scale easily. This
means that in order to support additional subscribers, we can add telephony
servers, storage capacity, voice circuits, and bandwidth to meet the demand
with little or no down time. Currently, we are at less then 25% capacity.
Whenever we exceed 40% capacity levels, we add more lines, servers, etc. to
handle the increased call volume. There is virtually no limit as to our
scalability options.
COA Network takes a very active stance towards keeping the
network as secure as possible. This starts with a very strong firewall policy
utilizing Check Point NG technology with extensive logging at all gateways.
|
|
|
|
|
|
|
|
Our
telephony servers are located on the internal network and can only be accessed
by users with administrative privileges. Network logins can only be created by
network super-administrators and are subject to a strict password policy. An
IP-based KVM system allows for strict control over access to servers by
designated users. Facility access, including data center entry, is managed
through an active key fob system controlled by network super-administrators.
All points of entry, as well as the data center, are under 24/7 video
surveillance and all key fob access is logged.
We
are pleased to announce that not one security breach has been encountered to
date.
|
An IP-based network monitoring system monitors the network 24/7 and
alerts our administrators of network and telephony connectivity, hardware, and
software issues through an escalated system of alerts (including e-mail and
cellular notifications).
Our DELL servers are under 24/7 same-day parts and labor warranty; in addition
our network administrators are trained to perform all hardware-related
maintenance. Many redundant systems are in place in order to minimize down time
from maintenance and repair. Since the software is created by in-house
developers, any software issues can quickly be resolved.
The databases supporting the system are backed up on a daily basis.
Using a combination of full, differential and transaction log backups, data can
be restored to any point in time within the last 2 months using MS SQL Server
toolset. Up-to-the-second selective transaction monitoring and rollback is also
available through a Log Explorer utility. All data is backed up to a
NetAppliance NAS device vertically scalable up to twenty terabytes, and a
redundant backup to tape media is made on a regular basis. Off-site backups of
mission-critical data are also performed.
Operating system, server and software upgrades are usually applied
between 4:00 AM and 7:00 AM EST to minimize service interruption for clients in
the continental US and Canada. System downtime during an upgrade is usually
measured in 15–30 minutes. All instances of planned system downtime are
communicated to clients via an appropriate area on the web-based login screen
at least 72 hours prior to the scheduled upgrade.
New features will be developed and incorporated to the system on an on-going
basis. Resolving bugs is a high priority for us; once a bug is identified, it
gets entered into our issue-tracking system and a developer immediately works
on resolving the issue. Once it has been fixed, a quality assurance engineer
will test to confirm the fix in our test servers. Then the fix is moved to our
production servers.
|
|
|
|
|
|
|
|
|
|