Schiller, Jason

Candidate Speech

Photo of Jason Schiller
Organization: Google
jason@schiller.net
http://www.google.com/ipv6/statistics.html

Jump to statements of support

Candidate Questionnaire

Bearing in mind that you run and serve as an individual, rather than representing an organization, is there an organizational affiliation you’d like to state?

Google

Are you available to serve the entirety of a 3-year term?

Yes.

Describe any limitations on your ability to travel to attend ARIN and ICANN meetings in person.

None.

Why do you want to serve on the NRO NC, and what goals do you want to accomplish?

I would like to serve on the NRO NC because there are a number of global issues surrounding IPv4 depletion and IPv6 adoption, and I believe it is important that we resolve these issues before the Internet suffers a tragedy of the commons. As a network engineer for a large content provider, I will bring level headed operations and engineering mentality to the table to in order to provide good policy which is in the best interest of the Internet as a whole.

At my core, I am an engineer and want to solve problems on their technical merit.

Some of the issues I would like to help resolve include:

As the ARIN free pool nears depletion, should ARIN allocate the smaller fragments that are left or will this overburden the routing system? What will be the impact of fragmentation of IP address markets on the routing system?

Post ARIN depletion, how should addresses be redistributed? Should the needs based system that the community has (up until now) deemed fair continue to be used? Does ARIN still have a stewardship role to play when they are no longer allocating or assigning IP number resources or is ARIN’s stewardship role most important when IP addresses become scarce (or prohibitively expensive on the open market)? Will an open market tend to concentrate IPv4 address in large and wealthy organizations? Does the open market have a different kind of fairness? Will organizations that have a real need for IP addresses be the only ones willing to purchase them, especially when their value will eventually be zero once wide spread IPv6 adoption occurs? Will it concentrate IPv4 addresses in organizations that derive the most revenue per IP address? If it does is that fair to services that derive little or no revenue per IP address?

How can we encourage the adoption of IPv6? How can we set the right balance between routability (fragmentation) versus sustainability versus administrative ease? How can we continue to support current IPv4 functionality while scaling to meet all of these potential IPv6 addresses?

What, if any, conflicts of interest might arise for you as an NRO NC member? Specifically, do you serve in what could be perceived as any Internet governance roles, provide any services directly or indirectly to ARIN, or represent any significant interest from the community?

None.

What is your record of serving the Internet community in the ARIN region?

I have served on the NRO NC, which performs the role of the ASO AC over the last six years. In that time I have served on the ICANN ASO AC nominating committee, and worked to craft many interview questions. I served on all Global Policy Proposal Facilitator Teams to shepherd global policy proposals.

I attended all ARIN meetings since ARIN XV (April 2005). I continue to be active in ARIN policy discussions both at the microphone in public policy meetings, and in informal meetings. I read PPML and have posted on occasion. I have authored policies and assisted others in crafting their policy proposals, and have been among a team of originators on a draft global draft policy, and an ARIN policy proposal. I presented at ARIN XX (October 2007) on the Implications of Global IPv4/IPv6 Routing Table Growth.

Starting this year, I am happy to be involved with Google’s web cast sponsorship.

I have attended and been a speaker at NANOG meetings since NANOG34 (May 2005).
Below is a list of presentations I have given:

– Inter-AS Traffic Engineering Case Studies as Requirements for IPv6 Multi-homing Solutions
(How IPv4 BGP TE is currently used as suggested requirements for an
IPv6 solution)
– Shim6: Network Operator Concerns
(What shim6 won’t do that can currently be accomplished with IPv4
Multi-homing)
– IPv6 Panel: Time for the Transition or Just More GOSIP?
(Is IPv6 worth the trouble?)
– Open issues with ipv6 routing/multi-homing [with Vince Fuller]
(Why we need a multi-homing solution that is not shim6)
– PANEL:Pragmatismv6: a Grown-up, Critical Examination of IPv6
(Discussion of operators realistically deploying IPv6 in a manor consistent with IPv4
operational practices vs. standards developers who idealistically want to avoid de-aggregation)
– PANEL: TEOTWAWKI: The end of the world, as we know it
(Discussion of operators about how IPv4 depletion will impact business as usual)
– BGP 101 and BGP 102 classes.

I attended several IETF meetings since IETF 61 (November 2004). I was a member of the Routing and Addressing directorate (RADIR). I have also been involved in the idr, grow, ipv6, v6ops, mboned, and pim working groups. I have made presentations to the GROW working group and the routing and addressing workshop (RAW).

What differentiates you as a candidate, or makes you uniquely suited to serve on the NRO NC?

I have served on the NRO NC which performs the role of the ASO AC for the last six years. I have a wide breadth of experience.

I have worked on medium and large sized Enterprise LANs for American University, Georgetown University, and the Georgetown University Medical Center. I have worked on a medium sized WAN environment at the corporate headquarters of MannorCare, a managed care company, that had 230 remote nursing facilities. I have worked in the day to day operations and the longer term engineering departments of UUNET (Verizon Business / MCI / WorldCom), a large carrier grade ISP. I am currently working at Google, a large content services provider.

I have a strong technical background, a wide variety of end-site, ISP, and CDN experience, small and large network experience, as well as LAN and WAN environments. I am active in the industry, and have a degree in international relations and philosophy which makes me uniquely suited for this role.

How do you propose to select an ICANN Board Member? Does there need to be any change in that selection process?

The most critical issues with the ICANN Board selection process is the time line. The process is overly rigid. The nomination phase is 60 days or longer. The comment phase starts at the end of the nomination phase, must be at least 30 days long, and must conclude when the interview phase concludes. The interview phase starts 30 days after the comment phase began, and must be no longer that 60 days.

In this 60 day window (or less interview phase), two or three rounds of interviews will be held. First a written interview will be conducted, which the candidate will have 10 days to complete. An ASO AC meeting must then be held (regularly the first wednesday of the month) for the NRO NC to review the written interviews and determine which candidates to move forward to the next stage. Depending on how large the candidate pool is, the NRO NC can then choose to simply have one round of telephonic interviews, one round of in-person interviews, or both a telephonic interview to further reduce the candidate pool followed by an in-person interview. The candidates must then be contacted, and the subsequent interviews scheduled. The interview will be conducted by an Interview Committee (IC) which is a representative sub-set of the NRO NC. They will generate and present a report to the entire NRO NC, during an ASO AC meeting, who will further reduce the candidate pool. This process happens twice in the case of having both a telephonic interview and an in-person interview. All within 60 days.

The most important part of the ICANN selection process is getting good candidates in the nomination phase. I am proud to say that we have had a very good slate of candidates from the ARIN service region for seat 9 in my time on the NRO NC:
– 2012: Thomas Eric Brunner-Williams, Martin Levy, Bill Manning, Ray Plzak
– 2009: Jody Newman, Ray Plzak, Rajesh Chharia, Suzanne Woolf, Barry Shein

In both those years someone from the ARIN region has been selected, which required seat 10 to be filled by a non-ARIN region candidate.

The next most important part of the selection is to try ask tough questions. I typically try to ascertain what the NRO NC collectively thinks are the weaknesses or concerns they have with each of the candidates and craft questions that get to the point of those weaknesses or concerns. It is then up to the candidate to convince us that there should be no concern, or confirm our collective suspicion. I invested a lot of effort in 2009 and 2012 in this process, and have been told by other NRO NC members that it contributed greatly to the process.

I remain committed to finding good candidates from the ARIN region and investing the time to craft and ask difficult questions. I have also begun a work on fixing our processes. I spearheaded an effort to revise our procedure to change our operating procedures to allow for eVoting when we could not get a 4/5ths majority to attend (never mind agree on) an ASO AC meeting. I have also significantly contributed to drafting a proposed procedure for ASO AC appointments to other positions. In the past, I have also contributed to revising the ASO AC’s ICANN Board selection process to make it less rigid (a vast improvement, but much work is still required).

Provide a brief biography of recent experience, associations, and affiliations relevant to serving on the NRO NC, including names of organizations, positions held, specific duties, and dates of service.

I have been with Google for the past two years as a network engineer in the production network operations group. I have done everything from turning up new peering interconnects, provided oncall support to network operations, and using BGP based traffic engineering in order to mitigate hot peering points, to certifying and deploying new hardware, as well as designing and implementing a new network to deliver traffic from the Olympics to YouTube for live streaming.

In the past year, my primary focus has been on IP number administration, coming up to speed as the chief numberista, and cleaning up, simplifying, and implementing best practices for BGP policy.

Prior to Google, I was with UUNET (Verizon Business / MCI / WorldCom) for over 13 years. My job responsibilities ranged from customer facing High Speed Install of 56K frame relay to OC-3 links, completing maintenance on the network, serving as a point of escalation, to long term architecture and engineering projects such as integrating AS701 and AS19262 (completed while I was at Google), architecting and implementing 3 phases of IPv6 (GRE over lay, 6PE, and native), transforming the BGP architecture to reduce one level of hierarchy while not impacting the number of routes, paths, or forwarding decisions, LatAm (AS14551) architecture and implementation, UUCAST (AS704) phase 2 network architecture and implementation, setting and maintaining global routing policy standards.

Prior to UUNIET I had combined three years experience in large scale LAN operations work at two universities, American University, and Georgetown University, and Georgetown University Medical center, as well as six months experience in a small scale WAN operations for ManorCare, a managed care company with 230 WAN connected remote nursing facilities.

In light of the ongoing transition from IPv4 to IPv6, do you have experience do you wish to share with the community that is noteworthy in terms of IPv6 adoption?

I am proud to be responsible for UUNET (Verizon Business) offering a global IPv6 Launch just one week after the IANA depletion was announced. My effort included developing the IPv6 numbering plan, architecting and deploying IPv6 in the UUNET North America network in three phases, GRE overlay, 6PE, and native, and native IPv6 in Europe and Asia.

I am proud to work for Google, a company that has not only embraced IPv6 in a real way, but has made significant progress in pushing the wider community toward IPv6. Admittedly much of this was had been completed prior to joining Google. I continue to work closely with Lorenzo, Erik, and Mike Joseph to advance IPv6. As numbers administrator for Google, I ensure all products and services who have a need for IP number resources, either support IPv6 or have a plan to do so. Additionally, I require a well formed and well documented IPv6 numbering plan for each such product or service.

In customer outreach for UUNET, I have been insistent that the community needs to understand that the deadline to support IPv6 is not the day before their individual organization runs out, but rather the day the first organization runs out of IPv4 and is forced to make the difficult decision of offering an IPv6-only solution, degraded IPv4 over Carrier Grade NAT (CGN), reclaim low margin uses of IPv4 to be repurposed, stop growing, or purchasing IPv4 space on the open market. If most organizations deployed dual-stack, then when an individual organization depletes their available IPv4 addresses, they could simply transition to IPv6-only for new products, services, or customers.

I understand that there is real cost to deploying IPv6 whether it be equipment upgrades, recertification efforts, complexity of deploying new code and new configuration, or training. This comes with no new products, no new services, no new capabilities, nor new revenue. As such a deferral not only saves money, but allows your competitors to spend their time, and risk their network health to fix issues introduced with IPv6. The problem with this approach is deploying IPv6 for a large provider is a likely a two year process. In my mind it is a requirement for networks, products, and services that lack IPv6 support, to make at least a minimal investment to deploy and offer a limited IPv6 service for all products, services, and configurations to insure that each organization has a working IPv6 solution. It is recommended to turn on IPv6 for all parts of the network and all products and services, and then expand that support to all parts that are already IPv6 capable. Once complete, the organization needs to survey the cost and time required to complete upgrades to fully support IPv6 and budget their time and money to start such upgrades before organizations start running out of IPv4.

This does not seem to have happened, and it is unlikely that the gap between IPv4 and IPv6 will be quickly bridged when organizations begin to run out of IPv4 addresses. This means even for organizations that have embraced dual-stack, they will be forced to commit to either purchasing IPv4 addresses, forcing traffic to the legacy IPv4 Internet for through an expensive and poorer performing CGN, or not making the legacy IPv4 Internet available to their new customers. These are tough choices when trying to maintain market share against competitors who have not yet depleted their store of IPv4 addresses.

What is your view of the existing bottom-up, self-governance model and structure of the Regional Internet Registry (RIR) system? Do you believe there are other models or structures that would work better?

The current bottom-up, open, and transparent RIR system is working well. This governance model allows for fair and equal participation of all segments the RIR community. It allows operators and IP managers to create, develop and adopt number resource policies that are technically sound, and in the best interest of their customers, their network, and the borderless, global Internet as a whole.

I believe there is no better model.

I continue to argue that the NRO NC should take a more active role in shepherding global policy and improving global communication and coordination. It takes a long time for the same text to be discussed in person in each RIR’s public policy meeting, modified, and agreed upon in all five regions. This is due to the timing of the meetings and lock-step nature of the global policy development process. Typically, global policy proposal text does not get modified until the results of a particular region’s public policy meeting suggest there are objections or concerns that need to be addressed, this can be a lengthy process.

It may be possible to shorten this timeline by cross-pollinating ideas and concerns from each of the regions as the discussion unfolds in each region. The NRO NC could form a Global Policy Proposal Facilitator Team (GPPFT) which could act as shepherds by summarizing concerns and arguments within their own region, and bringing those ideas to other regions. GPPFT members could also spur along conversations in their own regional mailing lists even if a meeting is not pending. In this way it may be possible to front load the conversation, and determine and address objections without having to wait for all five regions to have an in person meeting to get common agreed upon text. This could maximize the possibility that a global policy proposal or a globally coordinated policy proposal would pass on the first round of in-person RIR meetings.

 

Statements of Support

A forum for statements of support for this candidate is provided below via comments. Submit a comment to make your own statement of support for this candidate or simply view what others have posted. Submissions will be moderated before posting.