中文摘要：

海洋数据是一种典型的大数据，如何利用混合云存储架构存储海洋大数据是海洋数据管理面临的一个挑战．针对混合云存储架构中的关键问题-数据迁移，提出了海洋大数据的生命周期，并且基于此给出混合云存储中海洋大数据的迁移算法．在迁移算法中，将海洋数据的敏感度、数据访问频率、数据大小、数据时间长度等因素作为迁移因子．迁移算法兼顾考虑了数据存储容量、海洋数据本身的属性特征和数据访问过程中的动态变化．实验验证混合云存储模式能大大降低数据管理成本，同时，通过提出的迁移算法保证了数据的访问速度．

英文摘要：

Marine data is a typical big data. As the diversified marine data acquisition methods develop rapidly （remote sensing, GPS sensing, buoy monitoring, seabed monitoring, research ships and so on）, the marine data grows explosively. However, the management of big data is a sophisticated problem currently. Cloud storage is an effective way to manage the big data. As the big marine data is characterized by massive, multisource, uncertain and especially the sensitive, the conventional architecture of cloud storage is not suitable for it. The big marine data is managed on hybrid cloud storage platform based on its characteristics and applications. Then, how to take advantage of hybrid cloud storage to manage the big marine data becomes a challenge. Data migration is the key question in the hybrid cloud storage. To resolve it, the lifecycle of big marine data is put forward, and based on it, the migration algorithm for the hybrid cloud storage is proposed. In the migration algorithm, the marine data sensitivity, data access frequency, data time length, data size and so on, are provided as the migration factors. Migration algorithm considers the capacity of data storage, attributes characteristics of marine data as well as the dynamic changes in the process of data access. The experimental results show that hybrid cloud storage pattern reduces the costs of managing big marine data greatly. Meanwhile, the access speed of big marine data is ensured by the presented migration algorithm.