Network Management Policy

The Critical Hub's network consists of, but is not limited to, channel banks, hubs, switches, routers, servers, Local Area Networks at the equipment locations, and Wide Area Links connecting sites together consisting of microwave & fiber optic equipment, and the network management tools provided by the equipment manufacturer.

The System architecture is primarily constructed around an Internet Protocol based network. The network is composed of industry standard equipment, which also provides flexibility and a large variety of management & diagnostic tools. As a separate topic, the equipment configuration information will be included in this documentation.

The system network is complex and unusual problems may be difficult to identify and resolve. The system documentation will have to be kept up to date or will lose its value in supporting the system network. The system network is protected from other Critical Hub Networks data networks, and shall remain so. This is to protect the security and functionality of the system. If there is a connection to another data network, it shall be through an appropriately designed & maintained firewall. The components of the network shall be considered as "owned" by Critical Hub Networks, unless otherwise designated as a local component. In which case that component would be owned by the local unit of governance.

The individual owners (Collocation Customers) will then be responsible for the maintenance of the sites & equipment that they own. Agreements between the Owners and/or Maintenance Contractors are at each Customer's discretion, but the Owner is still ultimately responsible for their portion of the system. The Backbone system is structured on an integrated network; any infrastructure hardware and software upgrades or changes that may impact the system network will need reasonable discussion and subsequent approval by the Managers Group by a Maintenance Operations Protocol (MOP).

All maintenance work being scheduled that may affect the system and/or a local system performance shall be preceded by reasonable and appropriate notification to the Customers. The equipment configurations of the components of the network are documented at the Network Topology section. This is primarily for the purpose of maintenance, but also affects future planning. Critical Hub Networks generally provides best effort forwarding of all IP traffic. Exceptions to best effort forwarding are limited to the following:

The forwarding of traffic explicitly tagged as "less than best effort" by the end-user. Critical Hub offers a priority of traffic forwarding that is less than best effort to support researchers that choose a lower class of service when they transfer large datasets which are insensitive to loss or latency.

MPLS traffic engineering is sometimes used to differentially route Critical Hub's "managed high-speed service" traffic from commercial peering or commodity traffic for the purpose of load balancing traffic among backbone resources.

Critical Hub offers its network peers the option of using BGP community strings to signal to the Critical Hub backbone that traffic destined for a particular prefix should be discarded at the earliest opportunity. This mechanism is provided so that Critical Hub network peers are able to minimize the effects of an on-going network-based attack. In addition, network peers are provided with a range of traffic influencing BGP community values to control how their [the peer's] traffic is treated by the backbone.

The Critical Hub IP network is used exclusively for transit among regional networks and the public Internet (i.e., no users connect directly to the Critical Hub IP network). The Critical Hub IP network does not interfere with the transmission of content so long as the content is not harmful to the operation or security of the Critical Hub IP network and the networks with which it peers. Critical Hub is responsive to the lawful requests of law enforcement agencies, however given its role as a transit network, such requests are typically directed to the networks that serve end-users directly.

Critical Hub's standard practice is to maintain sufficient capacity to ensure loss-free packet transmission along its backbone facilities. Since the network provides two distinct types of service (e.g., public Internet access and a high performance managed service), different capacity planning strategies are employed for each service. The managed service maintains a high level (over 50%) of headroom to accommodate bursts of traffic associated with data intensive applications (e.g., HD video conferencing, large image transfers, etc.). Backbone facilities that support public Internet access are maintained with less headroom (i.e., 10%-30%).

Questions concerning Critical Hub network management policy should be directed to: noc@criticalhub.com