Network Planning Task Force

1
Network Planning Task Force

• Strategy Discussions

2
NPTF FY 07 Members

• Kayann McDonnell, Law
• Donna Milici, Nursing
• Dave Millar, ISC
• Michael Palladino, ISC (Chair)
• Jeff Fahnoe, Dental
• Mary Spada, VPUL
• Marilyn Spicer, College Houses
• Joseph Shannon, Div. of Finance
• Ira Winston, SEAS, SAS, Design
• Mark Aseltine/ Mike Lazenka, ISC
• Ken McCardle, Vet School
• Brian Doherty, SAS
• Richard Cardona, Annenberg
• Deirdre Woods/Bob Zarazowski, Wharton
• John Irwin, GSE

• Mary Alice Annecharico/Rod MacNeil, SOM
• Robin Beck, ISC
• Dave Carrol, Business Services
• Cathy DiBonaventura, School of Design
• Geoff Filinuk, ISC
• John Keane/ Grover McKenzie, Library
• Marilyn Jost, ISC
• Deke Kassabian /Melissa Muth, ISC
• Manuel Pena, Housing and Conference Services
• Mike Weaver, Budget Mgmt. Analysis
• Dominic Pasqualino, OAC
• James Kaylor, CCEB
• Helen Anderson, SEAS

3
Meeting Schedule FY 07

• Meetings 130-300pm, 3401 Walnut Street
• Fall Meetings
• Intake and Current Status Review August 21
• Agenda Setting Focus Group Planning September
  18
  
• Focus Group October 04
• Security Strategy Discussions October 16
• Focus Group October 17
• Network Strategy Discussions October 30
• Strategy Discussions November 6
• Final Meeting-Prioritization /Rate Setting
  November 20
  
• Focus Group Feedback December 4

4
Todays Agenda

• VoIP
• Wireless
• Authorization

5
VoIP

• Goal
• To convert 25,000 analog voice customers to
  Integrated Communications (VoIP, Voicemail, etc.)
  over the converged IP network with added
  functionality and lower costs in 5 years or
  less.

6
VoIP Current Status

• Developed flexible, cost-effective service based
  on open source VoIP.
  
• We have approximately 800 VoIP users
• SFS 50 new users in January
• Nursing Life Building - 50 users
• Vet Hill Pavilion Building 50 users
• Production-grade
• Redundant servers, gateways and PRIs
• Automated monitoring (server, network, PRIs)
• Single-line features, email/voice mail
  integration
  
• 911
• Architecture layer 2 Quality of Service,
  separate VLANs subnets
  
• Installation, Help Desk support billing fully
  operational.
  
• Customer and local support programs in place
• Local support provider training
• Reference materials
• www.upenn.edu/voice/voip/
• Deployed telephone directory services.
• Deployed web front-end to Integrated
  Communications Services.

7
VoIP Topology
8
Campus-wide Pilot Rollout
9
Penn iPhone Web Services
10
Penn iPhone Roadmap

• Phase 3.0
• First Multi-line Deployments
• Octel voice mail migration
• New Handset in Production
• Softphone Pilots - II
• 2000 phones deployed

• Phase 2.0
• Feature Release 2.0
• Softphone Pilots - I
• Centrex-Asterisk Pilot(s)
• ITSP/LNP Production
• New Handset Testing

• Phase 1.X
• IMAP Message Store
• Ring Group Pilots
• PiPS Enhancements
• Security Recommendations
• Handset Recommendations
• 600 phones deployed

• Phase 1.0
• VoIP Campus Pilot Begins
• Penn iPhone web portal
• Penn Directory lookups from iPhone

11
Current Development Projects

• Multi-line and ring-group support
• Using ITSPs (Internet Telephony Service
  Providers) for off campus calls
  
• Migrating Centrex users to our next generation
  voice mail
  
• Evaluating additional VoIP handsets
• Power Over Ethernet (PoE) network designs
• Support for emergency services and location
  (phone set moves)

12
Ring Group

• Ability to support multiple phone numbers on a
  single desk set
  
• Ability to have the same phone number ring on
  multiple sets, with the first user to pick-up
  getting the call
  
• These features being used internally within ISC
  today
  
• Management and data tracking being added to back
  end ISC systems to allow for full support at roll
  out
  
• Campus pilots to start in early 2007
• Roll out currently expected to happen by spring
  2007

13
ITSP Services

• Internet Telephony Service Providers (ITSPs) are
  an alternative to use of local PSTN gateways
  
• IP connection (dedicated and independent of
  Internet access) to a telephone service provider
  
• ITSP operates a large network (national,
  international) with many PSTN gateways
  
• Calls leave campus but go over IP until they
  reach a gateway near to the called party (the
  recipient)
  
• Could have multiple ITSPs and do least-cost
  routing
  
• Could maintain a combination of ITSP services and
  local gateways

14
Next Gen VMail for Centrex Users

• Todays Octel system works well, but is aging and
  is not able to interface with outside systems
  
• We have manual backup procedures for Octel.
  Restorations would take up to a day and we may
  not have everything restored (messages,
  greetings, and phone trees). 
• Todays Penn iPhone users are separated from
  todays Octel users
  
• Move to a single voice mail system
• Allows for the retirement of the Octel systems
• Brings some advantages of the Penn iPhone service
  to users still on Centrex
  
• New Voice Mail will have (always available,
  highly available) design with automated fail
  over).
  

15
Evaluating VoIP Handsets

• Possible criteria (abbreviated list)
• Support 802.1x, LLDP-MED, plus signaling and
  media security (TLS/SRTP)
  
• Working and supportable NAT traversal mechanisms
• Codecs for high quality sound (G722/E722)
• Mature presence management
• Integration with softphone, such as ability to
  import call logs, contact information shared
  between devices
  
• Backlit display with contrast adjustment
• Speaker phone, and Bluetooth wired headset
  support
  
• Ability to adjust display font and font size
• Some battery backup (15 minutes? enough to call
  911)

16
Power over Ethernet (PoE)

• Todays service requires that the desk set be
  powered locally (plugged in)
  
• PoE allows the network to provide power to the
  phone
  
• Some wiring closets have the space and power to
  support this. Some may not.
  
• Many other parts of the network would need power
  protection for this to be useful
  
• Could you live with a failover to your cell phone
  when the building power is down?
  
• Is the extra availability (99.99 vs 99.9) worth
  15-20 added to your bill?

17
Emergency Services Location

• Penn iPhone includes the ability to route
  511/911 calls, including support for location
  information
  
• We are building support for mobility and
  providing improved location information
  
• Three phases towards improvement
• Phase 1 - Fixed location phones
• Phase 2 - VoIP phones movable by LSP, using a web
  form for location reporting
  
• Phase 3 - Dynamic location updates

18
Collaboration within Higher-Ed

• Several major universities are using a very
  similar approach
  
• Sharing plans, code and best practices through
  monthly collaboration calls
  
• Open to participation by other groups seriously
  engaged in similar efforts
  
• Could result in a de facto standard for large
  university voice communications

19
Penn iPhone Estimated Cost Summary

• Assumptions
• Meridian Business Set one-time cost of 368 is
  depreciated over a 60-month period for this
  comparison
  
• 30 allocation is included
• Waived for first 7,000 pilot participants or end
  of FY 08

20
Voice (Misc.)

• CALEA (Communications Assistance for Law
  Enforcement Act)
  
• We continue to track developments
• We should know definitively about the level of
  compliance in January 07
  
• The potential cost impact for FY07 is likely to
  be about about 250k
  
• Current compliance date is in May 2007.
• Universal Service Reform Act of 2006 (HR 5072)
• We continue to track developments of the bill. On
  Apr 19, 2006, the bill was referred to the
  Subcommittee on Telecommunications and the
  Internet. It is in the first step in the
  legislative process. Introduced House bills go
  first to House committees that consider whether
  the bill should be presented to the House as a
  whole. Many of bills never make it out of
  committee.
• The fees are levied on anyone that
• Currently contributes to or receives universal
  service support
  
• Uses telephone numbers or Internet protocol
  addresses, or their functional equivalents or
  successors, to offer a service or a capability
  
• That provides or enables real-time 2-way voice
  communications
  
• In which the voice component is the primary
  function
  
• Offers for a fee, a service, that allows an end
  user to obtain access to a network that permits
  the end user to engage in electronic
  communications (including telecommunications)
  with the public.
• It is unlikely that there will be a cost impact
  in FY08.

21
Cellular Coverage

• Customers want improved cell coverage inside
  campus buildings
  
• The cellular companies-while willing to install
  outdoor towers/repeaters-will not fund inside
  coverage
  
• Should we increase phone charges for everyone to
  cover this cost?
  
• Should we imbed the cost in the CSF?
• Should we charge only the schools and centers
  that want it?

22
Wireless

• Vision
• Secure, seamless, cost-effective wireless
  connectivity for Penn community by June 2010.
  
• Drivers
• Smaller devices
• Mobility
• Customer expectation

23
Wireless (continued)

• Current Status
• We estimate that approximately 50 of campus has
  wireless connectivity
  
• 1000 ISC and school owned access points (APs)
• 500 APs added in College Houses and Sansom Place
  and 2 Greek houses
  
• Complete deployment of Wireless in College Houses
  and Sansom Place using 802.1X for authentication
  
• Current Strategy
• Should we change from our strategy of expanding
  the network AP by AP?
  
• Identifying the total cost for ResNet was
  effective at moving it forward
  
• Costs
• We estimate needing approximately 1000 more APs
  to get us to 100
  
• The one-time cost will be about 2M
• The ongoing annual costs will be 450k

24
Wireless (continued)

• What new strategies can we deploy to get to 100
  wireless connectivity?
  
• Move to 4 year depreciation on wired port
  electronics
  
• Reduce wired ports (6000 gets us 450K a year)
• Should we embed wireless infrastructure costs in
  the CSF?
  
• How much outdoor space is necessary?
• How about City Wireless with EarthLink?
• Penn program likely to be announced later in the
  fall

25
Central Authorization System

• Proposal for NPTF funding
• November 2006

26
Central AuthZ

• Build a central authorization system that could
  be utilized by applications across the
  University
  
• Utilize Penn Community data and school/center
  created lists to facilitate authorization
  decisions
  
• Allow Schools and Centers to build and reuse
  authorization information across applications
  
• Provide sophisticated group management
  capabilities, such as subgroups and composite
  groups, to support access management needs.
  

27
Benefits

• Facilitate consistent application of University
  business rules
  
• Managed through a common UI and standard API
• Streamline maintenance of authorization data
• Bring scattered redundant groups together for
  re-use
  
• Allow useful actions on these groups -- group
  math, group nesting, exclusion criteria
  
• Leverage Penn Community data for accurate, up to
  date authorization decisions
  
• Can leverage existing attribute information
• Support the creation of new groups
• By schools, departments, and individuals!
• Distributed/delegated model of control

28
Use Cases
29
Current Implementation Base

• More than 80 institutions are currently
  evaluating or implementing central authorization
  technology.
  
• Partial listing of where systems are in
  production or nearing a production cutover.
  
• Stanford
• Brown
• Cornell University
• Duke University
• National Cancer Institute (caBIG - Cancer
  BioInformatics Grid)
  
• South China University of Technology
• University of Bristol
• University of Chicago
• University of Kansas
• University of Southern California
• University of Wisconsin

30
Costs

• ISC would fund initial development and deployment
  (150,000)
  
• Proposal for 110,000 annual ongoing funding from
  NPTF (1 FTE HW facilities)