Internet storage service providers can store several petabytes of data and 
billions of files for thousands of geographically distributed users. This 
implies new challenges in scalability and manageability of large scale 
storage systems and services. The Active Object Store framework provides
a way for embedding executable methods in storage objects in order to 1) 
improve the performance by minimizing the network traffic and performing 
data processing tasks close to where the data reside and 2) improve the 
manageability of large storage systems by using self managing storage 
objects. This is a natural extension of  the Object-based Storage concept 
where storage objects contain not only data but also attributes. Explicit 
embedding of methods in objects allows users to perform data-intensive
operations like pattern searching close to where data blocks are stored 
in (for example ,  in client file cache, storage volume controller cache, 
or the storage device itself). Implicit embedding of methods by the 
storage system management software on the other hand can turn storage 
objects into self-managing, self-optimizing objects.  In this work, we 
design and implement a new storage system in which storage objects are 
made of data, attributes, and methods. More specifically, we define a new
API for extending the Lustre file system such that methods can be assigned 
to storage objects and executed.  We also study how this can enhance the 
scalability and  manageability of  the internet scale storage systems.