| 1 | = Minneapolis Meeting October 2007 = |
| 2 | |
| 3 | == Opt-In views == |
| 4 | GENI as an ISP: can be transparent - might be too transparent to not reveal to the user that they are part of an experiment |
| 5 | Generalized end-user services: issues and potential seti vs botnet |
| 6 | In-Network services: |
| 7 | different implications on the interface required |
| 8 | Retail vs Wholesale: |
| 9 | |
| 10 | == User Motivation == |
| 11 | faster than I2: |
| 12 | individual/institution bandwidth provide local limitations, protocol limitations |
| 13 | as easy to fix the performance problems for I2 as it is to opt-in to GENI |
| 14 | |
| 15 | Do you need to install software in order to opt-in? |
| 16 | |
| 17 | Value offered to user vs what pain and suffering is inflicted on the user: installing software - many dimensions to this (uses underlying stack, installs new stack, proxy, vm, etc) |
| 18 | traffic perturbation |
| 19 | |
| 20 | For wholesale, need a method to opt-out. |
| 21 | |
| 22 | Is there a history of using IRB process for networking |
| 23 | IMC2008 - Internet usage and privacy issues paper, will be posted |
| 24 | have generally had limited access |
| 25 | as a community we have not done this broadly |
| 26 | There are local constraints and capabilities which will also apply |
| 27 | |
| 28 | Could provide a template, where slices are available based on acceptable thresholds defining experiment attributes |
| 29 | |
| 30 | Need to bring in experts from other areas who can inform the strategy |
| 31 | |
| 32 | Another class of incentives - |
| 33 | Opt-In requirement on other NSF research - NSF could add incentives to other funded projects |
| 34 | |
| 35 | Scale - Implies big changes in research incentives, practices, personnel, and budgets |
| 36 | User characteristics are not real - especially wrt. attacks |
| 37 | Real users require real software |
| 38 | |
| 39 | Mobile users scale story the worst: 2B -> 4 |
| 40 | need to partner with existing service providers |
| 41 | Issue of understanding an economic model and motivation of the donors - homework that needs to be done |
| 42 | Costly to write "real" (robust) services - this doesn't just happen, costs real money, and needs to be supported after the creator graduates Have to Get real value if they are going to use the service |
| 43 | Potential solution: |
| 44 | Involve the central IT departments |
| 45 | Graduate to GPO |
| 46 | adopt firefox model - large distributed and professional development staff (variants - bsd, apache, etc) |
| 47 | |
| 48 | Most infrastructure which claims to support research has failed... wired is easier, for wireless need vendors for support |
| 49 | |
| 50 | Missing technologies |
| 51 | residential broadband... |
| 52 | |
| 53 | Realism: possible course of action |
| 54 | Remove or calibrated as a pillar |
| 55 | spend real money on clever packet generators |
| 56 | confront requirement for robust SW |
| 57 | Extend GENI to (all) desktops via sandbox technology |
| 58 | |
| 59 | Sometimes, the incentive is based on user perception, not necessarily reality |
| 60 | |
| 61 | Have we justified why we need real data? |
| 62 | |
| 63 | Can we specify/characterize what real data is? |
| 64 | |
| 65 | Don't want the observation/operation of the experimental parts to prevent transport as baseline service. |
| 66 | |
| 67 | Can spectrum lending be used as an incentive? |
| 68 | |
| 69 | == Organization == |
| 70 | incentives |
| 71 | technology (software/vm, etc) |
| 72 | legal/ethical/etc... |
| 73 | |
| 74 | == What is a user? == |
| 75 | Wholesale option - AT&T subscribes |
| 76 | Service Opt-In - CNN provides CNN' via GENI (wikipedia is being distributed via CODEEN) |
| 77 | |
| 78 | Research community does not have a history of developing the killer app, but we have provided the means for killer apps to be created. |
| 79 | |
| 80 | Access to resources they can't get otherwise |
| 81 | collection of graphics processors |
| 82 | distributed data storage |
| 83 | (expensive and large) |
| 84 | (imposes GENI as a veneer in front of something else - what is the motive for this service to allow GENI to impose itself in front of the service? The Places where GENI can impose itself are places above layer7) |
| 85 | |
| 86 | Layered motivation |
| 87 | Users want the innovative services that are built upon unique resources made available via the unique network (or services that rely only on the unique network) |
| 88 | This way we don't compete with our industry partners |
| 89 | |
| 90 | Should GENI (this WG) be involved in the formalization of IRB parameters and limits |
| 91 | NSF is interested in establishing a baseline |
| 92 | need to consider, exceptions (honey pots), local restrictions, international differences |
| 93 | what about local IRB processes, tools, etc |
| 94 | |
| 95 | 3 capabilities: |
| 96 | * Disclosure of who can gather data |
| 97 | * User can determine what data has been collected |
| 98 | * Correction - deletion of data at user request? |
| 99 | |
| 100 | Differentiate data collection for system administration from data collection for research |
| 101 | |
| 102 | Anonymization |
| 103 | Can provide a baseline outside of or streamlined IRB process |
| 104 | of limited use for some types of research and data |
| 105 | verifiability |
| 106 | |
| 107 | Taxonomy of opt-in classes |
| 108 | Taxonomy of protections and rights |
| 109 | |
| 110 | benchmark - Need a way to evaluate the opt-in stories in proposals (realize that some experiments can be successful w/o, but that the platform requires it for ultimate success) |
| 111 | |
| 112 | How much opt-in do we need? (how many users?) |
| 113 | |
| 114 | Is there a way to learn from previous test-bed opt-in failures? (impossible to opt-out and no users opted in) |
| 115 | |
| 116 | Traffic mirroring and synthetic load (use the users data, but don't perturb their packets) |
| 117 | Can also provide dual services and provide rollover capabilities |
| 118 | |
| 119 | Need to make sure the platform supports a continuum of reliability (99% - 5 9's) |
| 120 | Platform also needs to support monitor only geni-ized nodes |
| 121 | |
| 122 | How fast can you make the failover? |
| 123 | |
| 124 | Two opt-in problems |
| 125 | construction phase |
| 126 | need to include user opt-in here in order to ensure that they are there and supported post-construction |
| 127 | learning experience |
| 128 | limited opt-in support |
| 129 | needs to be a calculated risk - don't want to scare users |
| 130 | post-construction phase |
| 131 | |
| 132 | Opt-in strategy needs to be consistent with the maturity of the platform |
| 133 | |
| 134 | Open issues If users can bring in sensors or other devices which could involve other "users" |
| 135 | |
| 136 | Should the opt-in API be provided by the service or by the infrastructure (slice) |
| 137 | |
| 138 | – |
| 139 | Matt Sanders | Research Scientist | Academic and Research Technologies |
| 140 | Office of Information Technology | Georgia Institute of Technology |
| 141 | matt.sanders@oit.gatech.edu | 404-894-9107 | 105 Rich |