Transcript ppt - Common Solutions Group

In the Trenches
VOIP@Dartmouth
Peter Kiewit Computing Center
Dartmouth at a Glance
 Campus is roughly 1 mile square
 Geographically Rural
 Around 7000 extensions
 50% faculty/staff, 50% residential
 Around 120 buildings
Where we wereVoice Environment-2003
 Ericsson MD 110 PBX
 1500 digital phones
 4500 analog extensions
 Distributed Architecture-LIMs(line
interface modules)
 Main switchroom(35% of active extensions)
 34 LIM distributed among 5 locations
Where we areDual System Environment
 Cisco (Call Mgrs) solution integrated with
Legacy PBX(MD110)
 Faculty/Staff 100% VOIP-3100
extensions
 Resident Students(all analog) on legacy
PBX-3700 extensions
 Emergency services(analog) on Legacy
PBX-around 500 extensions.
Where we are going
100% VOIP Summer 2005
Faculty/Staff on full featured VOIP
phones
Residents on Analog over VOIP
Gateways
All greenfield installations on full featured
VOIP phones, faculty/staff or residential
How We Got Here
 Legacy PBX presented problems
 Costly upgrade/forklift on the horizon
 Upgrade path expensive with move to licensed price plan.
 Vendor essentially pulled out of North American
market, support difficult
 Telephony use at time of decision relatively
pedestrian(mostly analog)
 VM system no longer supported
 Revenue generation no longer a reality
How We Got Here
continued
 Budget Realities made convergence of
Voice/Video/Data operations attractive
 Cost to support separate operations for
Voice, Video and Data network high
 Newly Upgraded network made
convergence possible
 Aggressive Campus growth in 5 year plan
placed pressure on a “long term” direction
Business Case
 Upgrade/Replacement of existing System
to cost close to 2 million(pbx, copper
plant and Voicemail system)
 Greenfield installation costs began at
$50,000 for remote lim + voice
infrastructure
 Total Cost for Dartmouth VOIP
installation 1.4 million, system can
support 20,000 extensions.
Business Case
Continued
 Convergence of Voice/Video/Data
operations saved $400,000 annually
 Telecom, Network Services, CATV staffing in
2003- 23 positions, Compensation
$1,400,000
 Converged Network Services Current
Staffing- 12, compensation $970,000
 ROI on headcount savings alone is 4
years!!!
Conditions for Success
 Dartmouth Telephony/Network effort
highly centralized
 Legacy Telephony deployment not
feature intensive
 Most users did not understand how to
use the old Telephone and Voicemail
system.
 Old system cost prohibitive for
departments to take advantage of better
services
Common Objections we
were prepared for
 VOIP does not have the features of a
PBX
 VOIP is not as reliable
 My phone always worked when the
power went out
 You did not give us enough time to
prepare
Objections we were
not prepared for
 Can’t separate the “IP” in VOIP from the
Internet, confusion led to concerns over
privacy,spam and reliability.
 Users could be passionate about their
phones, especially cordless
VoIP Reality at Dartmouth
 Convergence and Integration
 Legacy forms and features preserved
 Just like the old system if you prefer
 Devices - telephones, laptops, hand-helds, etc.
 Features
 One number - many devices, or various groupings
 Dartmouth dial-tone from anywhere on the Internet
 Wireless platforms available across campus, voice
anywhere, anytime.
 Future: data, voice, and video integrated across
applications
 Unified messaging: E-mail <> Telephone, Text <>
Speech
 Video telephones
 Cell phone and wireless phone in-one
Path to Completion
Step 1-Build confidence in technology
 Pilot and Seed with Key personnel
 Install Core infrastructure with small production
deployments
 Get the “Money People” on board
 Package Technology as necessary for convergence
savings
 Get the right Deal!!!
 Get Support at the Executive Level(PAF)
 Public announcement of support from President’s office
Installation Plan
 Publicize the new technology
 Spend time on discovery-meeting with all
departments to review telephony needs
 Train Prior to installation
 Classroom training, not mandatory
 2 active systems allowed for us to work
at our pace
Faculty/Staff VOIP design
Deciding Factors
 Like-for-Like design
 Analog phones replaced with single line VOIP
sets
 Digital phones replaced with multi-line VOIP
sets
 All ancillary equipment replaced
(headsets…etc)
 POE only provided for VOIP communication
jacks
 60 minute UPS backup for all VOIP switches
Residential VOIP design
deciding factors-Why Analog?
 Most Students use cordless phones
today
 In VOIP model, cost is in the phones.
Unknown Abuse cost-recovery not
attractive
 Data Closets in Resident Halls lacking
 Power back-up and security an issue
 Life Safety Concerns
Life Safety-E911
 College E911 environment
 We supply PSAP with database of phone
numbers/locations
 We pulse out ANI to PSAP
 POE only supplied to “VOIP” data jacks to
discourage movement
 Residential extensions “hardwired” to insure
integrity of 911 database
 Cisco Emergency Responder being installed
Network Enhancements
for Production VOIP
 Core Network installed in a mesh
topology. Routers all have dual
supervisors/power supplies
 QoS enabled system-wide to insure
integrity of Voice Services
 Layer III routing enabled to VOIP edge
switches
 UPS backup for all closets and core
Gotcha’s from Cisco
 Beware of the upsell
 New technology for them causing “best
practice” design to change a fast pace
 QoS and layer III routing set-ups required
network upgrade as subscriptions for phones per
uplink changed dramatically
 Gateway boards required “Hybrid” mode in
routers
 Bottom line, get a contract guaranteeing that
design flaws are the vendor’s responsibility.
What went right
 The phones rang, message lights lit and
folks could make calls
 Logistics of cut went well, work at night
made cutover almost transparent
 Building by building cut allowed us to
change tactics on the fly.
 Users from old Analog world loved the
new features and ease of VOIP sets
 95% of campus thrilled
What went Wrong
 Not enough time spent in training to
explain subtle differences
 Reception features of new system not
adequate with software in GR at cutover.
 Lack of a consistent phone type led to
complaints
 Privacy panels on desks made
placement of phones difficult, extending
cutover
Lessons Learned
 Train staff prior to cutover or hire inhouse expertise.
 Make training mandatory and more
detailed
 Spend more time on “discovery” phase
with main call points(reception)
 Triage necessary with upgrades
Conclusion
 For years the promise of cost
savings has driven the advancement
of networks on college campuses.
Convergence of communications
services is finally delivering on that
promise, both technically and
fiscally. It is no longer a question of
if…but when.