Which ILM layout is ideal for large objects across three sites due to its efficiency and fault tolerance?

Which ILM layout is ideal for large objects across three sites due to its efficiency and fault tolerance?

• A. EC 4+2
• B. 3-copy replicated
• C. EC 2+1
• D. Balanced consistency

Answer: (A)

Erasure coding lets you store data as a set of data chunks plus parity chunks, so you can reconstruct the original object even if some chunks are lost. For large objects that must be available across three sites, you want both storage efficiency and strong fault tolerance. EC 4+2 distributes four data chunks and two parity chunks, meaning you can rebuild the object from any four of the six chunks. That lets you survive the loss of chunks from up to two sites without losing data, which provides solid fault tolerance across multi-site deployments. In terms of efficiency, this approach uses far less storage than keeping three full replicas for the same object, which is especially impactful for large objects. EC 2+1 would only tolerate a single failure, which isn’t enough when spanning three sites, and balanced consistency isn’t focused on reliability or storage efficiency across sites, so it doesn’t address the scenario as effectively.

Erasure coding lets you store data as a set of data chunks plus parity chunks, so you can reconstruct the original object even if some chunks are lost. For large objects that must be available across three sites, you want both storage efficiency and strong fault tolerance. EC 4+2 distributes four data chunks and two parity chunks, meaning you can rebuild the object from any four of the six chunks. That lets you survive the loss of chunks from up to two sites without losing data, which provides solid fault tolerance across multi-site deployments. In terms of efficiency, this approach uses far less storage than keeping three full replicas for the same object, which is especially impactful for large objects. EC 2+1 would only tolerate a single failure, which isn’t enough when spanning three sites, and balanced consistency isn’t focused on reliability or storage efficiency across sites, so it doesn’t address the scenario as effectively.