Plans and Preparation for Next Generation Networking

The Northern Tier Network Consortium How Do You Get to a Regional Optical Network When you ARE the Middle of Nowhere OR Northern Tier Networking - Not Exactly Broadband Special Version for the 2005 MT Government IT Conference Ray Ford - December 2005 AVP IT (CIO) - Professor of Computer Science The University of Montana MtGITC VERSION - DECEMBER 2005 CONTEXT Status Quo of State Gov’t/Ed Networking (SummitNet) Partnership between State/ITSD (numerous agency sites) and Montana University System (4 campuses of UM, 4 campuses of MSU), with roots going back 20+ years in telephone system operation, and with growing connections to local governments Implemented most prominently via a multi-year “backbone” contract with Qwest, but including other elements and commercial partners as well Qwest contract due to end soon has partners thinking “what next after Qwest contract?” Note: For various reasons state-wide K-12, community colleges, and tribal colleges are not full SummitNet partners State Funding Cycle - Next Biennium: Revision to State IT Strategic Plan in progress, next cycle of state agency plan submission and approval upcoming soon (Feb - June 2006) MUS Funding Cycle - Next Biennium: Revision of MUS Strategic Plan done (includes IT elements), consideration of proposals (including IT elements) by MUS and Board of Regents in progress with March BoR as target Northern Tier Network Consortium: UM, MSU, and ITSD are all members Northern Tier NSF Planning Grant: NT members ID, MT, ND, SD received $200K for next generation network planning -- activity starting NOW THE USA NORTHERN TIER - MOTIVATION If this was 1803-1806 We’d be Lewis & Clark without a map, hoping that the Missouri River system goes all the way across the Northern Tier If this was 1870 We’d be looking at a map trying to assure that railroad lines go into and across the Northern Tier If this was 1955 We’d be looking at a map trying to assure that the interstate highway system goes into and across the Northern Tier But because this is 2005 We’re looking at the map trying to assure that next generation networking goes into and across the Northern Tier [Apologies to ME, NH, VT, NY, MI, …, but “northern tier” = “northwestern tier”] THE USA NORTHERN TIER - THE CHALLENGES “Our region has a lot of dirt between light bulbs.” Attributed to Conrad Burns, US Senator/Montana In a world where network costs tend to be directly proportional to physical distance and network cost models are highly sensitive to the number of network users, having relatively few users spaced relatively far apart presents real physical and fiscal barriers of two major types Connecting to aggregation points: reaching commodity points (Seattle, Minneapolis, Denver) Connecting end service points: extending appropriate bandwidth to current/potential customers THE NORTHERN TIER NETWORK The details via WHO, WHAT, HOW, WHY, WHERE, WHEN, … WHO? NORTHERN TIER NETWORK CONSORTIUM “www.ntnc.org” HISTORY Formed Spring 2003 by reps from WA, MT, ND, SD, MN, WI -- later joined by ID, IA, NB [AK?] - Charter members: major universities across the region, with GigaPoPs at each end - Joined by other academic, non-profit, and state/federal government entities - Sponsored as a “project” by Internet2 Goal 1 Upgrade “east to Chicago” and “west to Seattle” links, extending and connecting state networks Goal 2 Build “golden spike” link to complete Seattle/Chicago throughway Form Build, combine, coordinate state networks (in/into the region) and extend the national network footprint (across the region) • 6 WHO? NORTHERN TIER CONSTITUENCIES AND FOCUS Ford’s Analysis and Claims • • • • • Higher Education: Connections between state sites and to WORLD are equal #1 priorities State/Local Government: Connection between state sites is #1, to WORLD has been a distant #2 -- importance of to WORLD is growing, as is promotion of to consumers Commercial/Non-Teleco: Connection between company sites (typically out of state) is #1, to WORLD (customers) is close #2 Commercial/Teleco: Connection to aggregation points and service points are equal #1 Consumers: Connection to WORLD is #1 -- consumers typically don’t distinguish between WORLD vs. state services, nor “connection to the home” vs. access to aggregation points (at least not yet, but its coming -- look at cell phone advertising). The NTNC focus has been on aggregation, moving bits more rapidly/cheaply between NT states and aggregation points. This aligns very well with Higher Ed’s interests, not so well with traditional state/local gov’t interests, very well with commercial interests, and very well (I claim) with consumer interests. WHAT? NEXT GENERATION (“XTREME”) CONNECTIVITY Not Consumer Broadband -- The Equivalent for High End Users or Providers Consumer Bandwidth - Entry Point (dial up): <= 56 Kbps - High Speed (Broad Band): 128K-500Kbps Traditional Commercial Bandwidth - Old Low End (ISDN): 96Kbps - Now Low End (T1’s): 1.5 Mbps - Old High End (DS3’s): 45Mbps - Now High End (OC’s): OC-N = N * 55Mbps (e.g., OC-48 = 48 * 55Mbps = 2.4Gbps) Xtreme Bandwidth (lambdas on fiber pairs) - Yesterday: 1 to 45 lambdas, at 10Gbps each - Today: 1 to 45 lambdas at 40Gbps each HOW? LIGHT FIBER WITH XTREME EQUIPMENT From National Aggregation Center(s) to/through Local Trunks Cost Components - In General - Build long haul fiber infrastructure (fiber + huts) - Assure local access and local support (“remote hands” arrangements) Both of the above done on major routes, need to be done on others Xtreme Specific: Upgrade equipment to light the fiber with Xtreme equipment Today’s Reality Networks based on OC48 (2.5Gbps) equipment, designed to be aggregated/sold in units of 1.5-45Mbps Xtreme Reality Moving to a backbone based on multiple 10/40Gbps lambdas is technically possible, but takes a major capital investment (new business model, viable business case) HOW? XTREME NETWORK OWNERSHIP AND OPERATION New or Old Operational Models Who Would Build, Own, and Operate? - Option 1: A consortium builds, owns, and operates the infrastructure [lighting dark fiber] - Option 2: A consortium buys services from a vendor [service acquisition] - Option 3: A consortium works with vendor(s) to co-build, co-own, and cooperate the infrastructure [shared ownership, “dim fiber” model] NOTE #3 is most likely the preferred option in the current telecomm climate WHY? TO SUPPORT NEXT GENERATION APPLICATIONS AND AGGREGATION The goal is to - Scale aggregation bandwidth (consumer, commercial, education, government) dramatically beyond what is now possible - Enhance current bandwidth-constrained applications - Enable novel applications that inherently require more (xtreme) bandwidth The Goal is NOT to make existing (non real time) applications like email work faster -these may be constrained by connectivity purchased at the end point, but are not constrained by trunk network capacity or technology available for end point connection The Goal is NOT to make fast Internet access a public utility or a “citizen entitlement” (though some would advocate for these) NOTE Assuring that “available end point technology” is ACTUALLY available at all end points is partially an aggregation issue, but also involves additional issues of capital investment, private vs. public utility policy, etc. WHY? TO SUPPORT NEXT GENERATION AGGREGATION Gross Model of Business/Consumer Aggregation Assuming that customers buy connectivity at a given volume but on the average utilize only a certain percentage, here are some simple computations about the number of users different trunk infrastructures will support INFRASTRUCTURE 2.5G 10G 40G 500,000 2M 8M 50,000 200,000 800,000 10,000 40,000 160,000 250 1000 4000 125 500 2000 5 20 80 0.5 2 8 Consumer-50K Consumer-500K Consumer-500K Pub/Comm’l-50Mb Pub/Comm’l-100Mb Pub/Comm’l-1Gb Pub/Comm’l-10Gb Util 10% 10% 50% 20% 20% 50% 50% N * 40G N * 8M N * 800K N * 160K N * 4K N * 2K N * 80 N*8 Montana CANNOT achieve its objectives for consumer connectivity or expansion of high tech industry without a significant upgrade in aggregate connectivity WHY? TO ENHANCE CURRENT BANDWIDTH CONSTRAINED APPLICATIONS EXAMPLE - REMOTE PRESENCE Current “standard” (H320/323) Single A/V stream -- 384/768K EG: MetNet, VisionNet, … Current “high end” (access grid node) Multiple concurrent A/V streams -- 1Mb to 10Mb to … (depending on number of streams) EG: AGN’s at UM, MSU-Bozeman, and MtTech routinely connect 3-9 streams from MT, region, country, world Vision for Tomorrow: R&D in progress on improved 3-D-like display, perhaps even real time holographic display -uses LOTS (G’s, not K’s or M’s) of bandwidth WHY? TO ENABLE NEXT GENERATION APPLICATIONS Example - Remote Instrumentation Consider equipment controlled by a high resolution (HDTV-like) display, e.g., 1000 x 1000 pixel display, 30 frames/sec , 36 bits/pixel color) Bandwidth Requirement (10**3)*(10**3)*30*36 = 1Gbps OBSERVATIONS/VISION Cable limits have forced equipment console to be co-located with the equipment The “console” could be a Web browser accessible anywhere the network can sustain Xtreme (>=1Gbps) performance National Cyberinfrastructure Plan: Provide research/education access to equipment by implementing this vision WHEN? WHEN DO WE NEED XTREME NETWORKING? YESTERDAY OR As Soon As Possible WHEN? CURRENT PHASED ROLLOUT VISION MULTI-STEP IMPACT PLAN 0. NSF $200K planning grant awarded to ID, MT, ND, SD -- activity underway NOW 1-A. MUS plans to build out Seattle to MUS-I90/94 (including main SummitNet trunk) -- proposals for federal R&D funds are pending NOW 1-B. Hope to combine new SummitNet contract with 1-A, plus add new “loops” through NE & NW MT 1-C. Assure that 1-A & 1-B provide incentives for vendors to reach new markets, provide new services 2. Opportunity-Based Extension: All concerned look for follow on opportunities to extend to libraries, K-12, tribal colleges, local government, major health care facilities, pockets of business development, and consumers KEY PRINCIPLE Public entities cooperate, not compete, with commercial providers to develop the new network market. MORE DETAILS IN SPECIAL OPPORTUNITIES FOR MUS Internet2 and National Lambda Rail Commercial Network Traffic travels - via commodity internet -- crowded and as a result with unpredictable speed - via private/virtual private networks -dedicated service to internal sites at various speeds is available (at premium prices) University Network Traffic travels - via commodity internet to non-university endpoints and VPNs to internal sites - via subsidized/restricted Internet2 network to member university endpoints - via subsidized/restricted extreme NLR network between NLR member endpoints MORE DETAILS ON CURRENT MUS ACTIVITY MUS Connectivity Current and Planned University Network Traffic - Current - Campus/Campus: via SummitNet (VPN) at various speeds (1.5-45Mbps) - Campus/State: via SummitNet (VPN) - Commodity Internet: UM - 45Mbps to VisionNet/Main, MSU - 45 Mbps to Qwest - Internet2: UM & MSU - 45Mbps each service to Pacific NW Gigapop in Seattle (each using DS3 transport from TA/360 Networks) New/In Development - Lariat Project: grant funded 2.5Gbps service from PNWGP to MSU-Bozeman - Pending Projects: proposals for multi-10Gbps service from PNWGP to Spokane (partners) to Missoula (UM and partners) to Bozeman (MSU and partners) to … MORE ON HOW? GETTING ON I2/NLR MAPS, BRANCHING OUT ON MT MAP Per Interstate Highway Analogy - Co-build expressway into MT -- MUS and national R&E partners - Co-build trunk expressway in MT -- ITSD, MUS, and state partners (ITSD, MUS, vendors) -- an extension of SummitNet - Co-build richer set of access roads (loops) between expressway and other sites -- an extension of SummitNet - Leverage similar development in nearby states (especial ND) to co-build expressway across MT and the whole northern tier Additional Information: NTNC Distances Or, Consider Linear Distance Route Drive Dist. Seattle/Spokane 282 Spokane/Missoula 199 Missoula/Bozeman 200 Bozeman/Billings 145 Billings/Wibaux 245 Wibaux/Bismark 165 Bismark/Fargo 200 Fargo/Minneapolis 240 Minneapolis/Madison 272 Madison/Milwaukee 77 Milwaukee/Chicago 92 Total - Seattle/Chicago run Fiber Miles 379 199 203 151 300 *** 170 *** 191 250 280 117 104 2353 miles Additional Information: NTNC Density Harsh Realities of Network Ownership State Area People S.M./P [Alaska 570.K 0.6M 0.910 ] Wyoming 97.1K 0.5M 0.197 Montana 145.6K 0.9M 0.161 N.Dakota 69.0K 0.6M 0.107 S.Dakota 75.9K 0.8M 0.100 Idaho 82.8K 1.3M 0.063 Nebraska 76.9K 1.7M 0.044 --------------------------------------------------------------Iowa 55.9K 2.9M 0.019 Minnesota 79.6K 4.9M 0.016 Washington 66.6K 5.9M 0.011 Wisconsin 54.3K 5.4M 0.010 Contrast With Texas 261.9K California 156.0K Massachusetts 7.8K Rhode Island 1.0K 20.9M 33.9M 6.3M 1.0M 0.013 0.005 0.001 0.001 Source: 2002 Rand McNally Road Atlas Additional Information: Money and Costs General Fiber is available under certain conditions from Seattle to Chicago (ATT/SURA) and “secured” (PNWGP) from Seattle to Billings Straw Proposal NLR, Fall 2004, $20M over 5 years to build and operate Seattle-Chicago link with P-O-P’s along the route Seed Funding NIH/Lariat grant to MSU Planning ND/SD/ID/MT have just received a $200K planning grant from the National Science Foundation Funding ND/SD/ID/MT are currently seeking state and/or congressional funding for their regional pieces Additional Information: R&E Network Politics Exciting Network Merger Times - I2 and NLR are merging - I2’s Network Policy & Planning Advisory Committee (NPPAC) is a major force determining the future - NPPAC includes the CIOs from 15 prominent universities - Three NPPAC reps are from the NTNC -UWa, UWi, and UMt -- me (see if you can find me in the pictures below) NTNC Summary We’re not connected yet, but - We are active in promoting a vision of collaboration across the region, across organization types - We are active in looking for money and partners for state and regional efforts - We are active in looking for money and partners for inter-state/national efforts - We are now starting to look at design and planning Ideally we’ll use our NSF grant to do the planning, get state by state funding and get ready to build, then attract national interest to build it “right” with all the right partners Stay tuned … NTNC - Questions?