Exercise 1: Classifying Storage Needs — Possible Solution
====================================================================

(a) An OS boot disk for a VM.
    -> BLOCK STORAGE. Per the chapter, block storage is specifically
    "used for OS boot volumes" -- it needs to behave like a raw,
    mounted disk with low-latency, random-access I/O, which is exactly
    what an operating system needs to boot and run from. Object
    storage isn't mountable as a bootable filesystem at all, and file
    storage is meant for shared multi-VM access, not a single VM's own
    boot volume.

(b) Millions of small log files that must be retained for compliance
    for 7 years.
    -> OBJECT STORAGE. This matches the chapter's own "typical use"
    for object storage directly ("backups, static assets, logs, data
    lakes") -- large numbers of individual files, accessed
    occasionally rather than needing low-latency random access, and
    needing to scale to "millions" without practical limit, which
    object storage is specifically built for. The 7-year compliance
    retention period is also a strong signal to combine this with
    Chapter 4's own lifecycle-policy material: these logs are an ideal
    candidate to automatically transition into cheaper archive tiers
    over time as they age, rather than staying in the (more expensive)
    standard tier indefinitely.

(c) A shared directory accessed simultaneously by 10 VMs.
    -> FILE STORAGE. This is precisely the gap the chapter identifies
    neither block nor object storage solves on its own -- "shared,
    simultaneous access from multiple VMs at once." Block storage
    attaches to one VM at a time in the standard configuration, and
    object storage isn't a mountable shared filesystem. File storage
    (EFS/Azure Files/Filestore) exists specifically for this exact
    scenario.

WHY THIS WORKS AS AN ANSWER
------------------------------
Each classification traces back to the specific defining property the
chapter attaches to each storage type -- boot-and-database access
patterns for block, massive-scale/infrequent-access files for object,
and simultaneous multi-VM sharing for file storage -- rather than
guessing based on the word "storage" alone.