A. General Description of the Application
  1. TLD String(s) Requested.
  2. Category.
    General Purpose, Restricted Content.

    DotKids, Inc. (“DotKids”) requests the .kids TLD. The application implicitly attempts to restrict content across the TLD and targets a more narrow group of registrants. The targeted end user base, kids and their families, is very broad. Accordingly, DotKids qualifies for the general purpose category and restricted content group.
  3. Sponsor, Registry Operator and Subcontractor.
    a. Sponsor. Unsponsored Application.
    b. Registry Operator. DotKids was incorporated on September 29, 2000 as a Delaware corporation with its principal address in Rosemont, Illinois. While DotKids is a newly formed company, it claims to bring together over 60 years of relevant experiences in establishing successful business operations. DotKids intends to utilize SARAF Software Solutions, Inc ("SARAF") for the software development of the .kids registry system and all technical support and future enhancements to the system and Comdisco to provide the data center and daily operations of the registry along with the backup and disaster-recovery backup sites.
    c. Subcontractor. SARAF claims to have over five years of Internet application development experience. SARAF claims its focus on Internet backend applications with a niche in domain registration systems and backend infrastructures allows it to offer robust e-solutions to e-business needs. Comdisco, located in Rosemont, Illinois, claims to have over 30 years of experience and offers a complete suite of information technology services including business continuity, Web services, network services, and IT Control and Predictability Solutions SM. Comdisco also claims to deliver flexible equipment solutions for key vertical industries and a unique blend of technology services for education and healthcare organizations. Finally, Comdisco claims to provide disaster recovery to about 75% of the fortune 500 companies.
  4. Registry-Registrar Model.
    DotKids intends to only deal with ICANN accredited registrars.

B. Technical Review
  1. Summary Description of Proposal.
    DotKids proposes an unsponsored TLD to act as a parallel to .com aimed specifically at children. Enforcement of the registry restrictions would come not through filtering registrants but by requiring contractual commitment to standards.
  2. Support of the Business Plan by the Technical Plan.
    a. Total Capacity. The application does not describe proposed equipment, makes minimal load projections and does not characterize what, if any, changes in load will result from the proposed changes in registry data and registry-registrar interface protocols. Clearly Comdisco has the infrastructure and technical skills to deploy a high capacity system. It is unclear whether the software provided by SARAF is similarly scalable.
    b. Projected Growth Rate. The proposal relies on Comdisco scalability to handle growth. It would be important to understand whether the SARAF software can take advantage of Comdisco’s scalability.
    c. Startup Period. DotKids projects a slow startup, with a Sunrise period, making it unnecessary to make extraordinary preparations for high volumes at startup. This is an assumption that needs to be more thoroughly justified.
    d. Fault Tolerance. DotKids proposes to use the Comdisco Single Site High Availability option. No mechanisms were described as to how this specific application would use this option. While the actual uptime value of the Multi-Site High Availability option is small, it does seem like an option that the applicant should pursue.
    e. Security. Standard security models are proposed.
  3. Summary of Relevant Experience.
    Other than the Sarafs, who are associated with the contractor SARAF Software Solutions, DotKids management has limited direct experience running Registry operations or global non-stop network systems. Many of the individuals on the various advisory boards have been involved for a long time with issues related to children using the Internet.

    The subcontractor responsible for hardware and networking, Comdisco, has extensive experience with high availability, high performance computer based systems.

    The subcontractor responsible for software, SARAF Software Solutions, has prior experience in the design of the RRP protocol and the systems and processes currently used by NSI.
  4. Apparent Implementation Risks.
    The SARAF Software Solutions product RegyStar has not been used in high volume applications. Parts of it may still be in development. The company is relatively new, making it difficult to judge the quality and stability of its work.

    The proposal is for the main servers to reside at a single site, although Comdisco also offers a multisite high availability solution, which would seem to be more preferable for a global system such as this one. In general, operating successfully in a multi-site model requires careful design. The proposal does not speak to the geographic distribution of root name servers.

    The content rating information is retained in the registry and apparently describes the rating(s) for the entire site. The application did not specify how this information would be made available to browsers. Also, no estimate was made of how retaining this new piece of information in the registry might affect traffic levels.

    More broadly, it is unclear how the Rating Systems would be used in practice, how the information might flow to third level domains and below, how many rating systems would be supported and how registrars would interact with raters to ensure that the information residing in the registry would be accurate. As errors in reflecting rating information might have severe commercial consequences, the proposal also should show how such risks might be minimized through technical means.

    The application states on page 32 “The .kids Registry model is much more complex than current model [sic] (NSI) used in the registration of .com, .org and .net.” Introducing a more complex model requires that registrars invest in software development. Depending on the complexity of the development and maintenance of that software, registrars may find this requirement to be a technical risk.
  5. Available of Human, Operational and Technical Resources to Cope with Unexpected Events.
    Comdisco has the technical resources to respond to most technical emergencies, as well as the technical knowledge about how to design to avoid the most obvious surprises.

    SARAF presumably is well connected in the Internet technical community. The company appears to have a small staff available for software development; responsibility for operational problems was not clear.
  6. Advancing the State of the Art.
    The proposal envisions a “thick” registry, similar to the CORE model. Apparently this is in part to support the retention of Ratings information in the registry.

    As the proposed registry software, RegyStar, is third party software, it may provide a suitable platform for future TLD Registry operators. This might be a useful vehicle for getting the software debugged in production environments.
  7. Other Comments.
    The introduction of Ratings into the registry might be a useful model for other special purpose TLDs. Unfortunately, the application is silent on how that information is made available to running software.

C. Business Review
  1. Applicant’s Representations.
    DotKids was established specifically for this unrestricted TLD application and has no operational history. Its mission is to incorporate multiple processes to address the issue surrounding minors online. The company has seven employees and eight Advisory Board Members. It has a formal alliances with SARAF Software Solutions, an Internet solutions company with a niche focus in domain registration systems, and Comdisco.

    The revenue model includes a $10,000 fee from all ICANN-accredited registrars. During the initial rollout period, DotKids will charge $48 per registration during the first month, $36 during the second month, $24 during the third month, and $12 beginning in the fourth month. The target market is defined as websites for children, teenagers, and parents and will be focused on education, entertainment, health, sports, business and research. The application notes that there are 25 million 2-17 year olds online in the United States, 68% of teenagers are currently online, and 62% of children (2-12 year olds) will be online by 2005.
  2. ICANN’s Evaluation.
    The strengths of this application lie in (1) the almost universal desire to ensure children have safe Internet experiences, and (2) the domain registration experience of an alliance partner. The application’s weaknesses include irregularities in the pro-forma projections and the current lack of capital funding. Overall, there are other applications in this category that are stronger from a business plan perspective.

D. Summary of Public Comments
  1. Number of Comments.
  2. Support for the Application.
    By making the rating system voluntary, parents and families are given a role in making decisions.

    The proposal will allow parents to restrict children’s access to inappropriate content.
  3. Opposition to Application.
    Filtering technology is imperfect and thus the .kids TLD will not prevent children from being able to access inappropriate content on-line.

    A question was raised regarding the qualifications of Dotkids to follow through on the proposal.

    The .kids proposal would not rely on objective criteria for categorizing content and thus the registry would be able to restrict content based on its value judgments. (Same comment posted for other .kids proposals).