Today we continue discussing the best practice from storage engineering:

455) The use of a virtual machine image as a storage artifact only  highlights the use of large files in storage. They are usually saved on the datastore in the datacenter but nothing prevents the end user owning he machine take periodic backups of the vm image with tools like duplicity. These files can then be stashed in storage products like object storage. The ability of S3 to take on multi-part upload eases the use of large files.

456) The use of large files helps test most bookkeeping associated with the logic that depends on the size of the storage  artifact. While performance optimizations remove redundant operations in different layers to streamline a use case, the unoptimized code path is better tested with large files.

457) In the next few sections, we cover some of the testing associated with a storage product. The use of large number of small data files and a small number of large data files serves the most common case of data ingested by a storage product. However, duplicates, order and attributes also matter. Latency and throughput are also measured with their data transfer.

458) Cluster based topology testing differs significantly from peer-to-peer networking-based topology testing. One represents the capability and the other represents distribution. The tests have to articulate different loads for each.

459) The testing of software layers is achieved with simulation of lower layers. However, integration testing is closer to real life scenarios. Specifically, the testing of data corruption, unavailability or loss is critical to the storage product

460) The use of injectors, proxies, man-in-the-middle test networking aspect but storage is more concerned with temporary and permanent outages, specific numbers associated with minimum and maximum limits and inefficiencies when the limits are exceeded.

#algorithm
MST-Prim
// this grows a tree
A = null
for each vertex v in G, initialize the key and the parent
Initialize a min-priority queue Q with vertices
while the Queue is not empty
       extract the vertex with the minimum edge distance connecting it to the tree
       for each adjacencies v of this vertex u
              set the key to the weight(u,v) and parent

Print Fibonacci using tail recursion:
uint GetTailRecursiveFibonacci(uint n, uint a = 0, uint b = 1)
{
    if (n == 0)
        return a;
    if (n == 1)
        return b;
    return GetTailRecursiveFibonacci(n-1, b, a+b);
}