Thus, by storing k data units in each of different territories, especially where a network, power, or other failure will affect only one of these territories, end user 1095 will typically have more reliably access to the information in case of a failure (including maintenance) of a hive node.