Revision 28c9b5b0 docs/archipelago.rst
b/docs/archipelago.rst | ||
---|---|---|
3 | 3 |
Volume Service (archipelago) |
4 | 4 |
^^^^^^^^^^^^^^^^^^^^^^^^^^^^ |
5 | 5 |
|
6 |
`Coming Soon ...` |
|
6 |
Introduction |
|
7 |
============ |
|
8 |
|
|
9 |
Every Volume inside a VM can be thought of as a linearly addressable set of |
|
10 |
fixed-size blocks. The storage of the actual blocks is orthogonal to the task of |
|
11 |
exposing a single block device for use by each VM. Bridging the gap between the |
|
12 |
VMs performing random access to Volumes and the storage of actual blocks is |
|
13 |
Archipelago: a custom storage handling layer which handled volumes as set of |
|
14 |
distinct blocks in the backend, a process we call volume composition. For the |
|
15 |
actual storage of blocks we are currently experimenting with RADOS, the |
|
16 |
distributed object store underlying the Ceph parallel filesystem, to solve the |
|
17 |
problem of reliable, fault-tolerant object storage through replication on |
|
18 |
multiple storage nodes. Archipelago itself is agnostic to the actual block |
|
19 |
storage backend. |
|
20 |
|
|
21 |
Archipelago is under active development and will be available soon. |
|
22 |
|
Also available in: Unified diff