The pet-photo mountain

The storage account

An Azure storage account is the container for your Azure Storage data: blobs, files, queues and tables all live inside one. It gives that data a unique namespace you can reach over Hypertext Transfer Protocol (HTTP) or HTTP Secure (HTTPS) from anywhere in the world, and it's where you pick settings that apply to everything inside it, such as performance, redundancy and the default access tier. Pawsome Pets keeps everything in one account called pawsomepetsdata.

Sam: One account for everything. Bold choice.

Dev: Not wrong, just crowded. Let's see what's in it.

Five services, five jobs

Azure Storage has several services. Pick by the shape of the data and how it's used:

• Azure Blob Storage: huge amounts of unstructured data (photos, video, backups, logs), reached by web address (URL)
• Azure Files: fully managed file shares that computers mount with Server Message Block (SMB) or Network File System (NFS)
• Azure Queue Storage: large numbers of messages, up to 64 KB each, so parts of an app can pass work along
• Azure Table Storage: NoSQL (non-relational) key-attribute data in tables; structured, schemaless and inexpensive
• Azure Disk Storage: managed disks, the block-level volumes that Azure virtual machines (VMs) use as their drives

Exam tip Objects = Blob. Shared drive = Files. Messages = Queue. Key-attribute data = Table. VM drive = Disk.

Blobs up close

Blob stands for binary large object. Blobs sit in containers inside the account, a bit like folders. There are three blob types: block blobs for everyday text and binary files such as photos, append blobs for data that only gets added to, like logs, and page blobs for random read/write files such as the virtual hard disk (VHD) files behind VM disks. Every blob has its own URL, for example https://pawsomepetsdata.blob.core.windows.net/photos/biscuit.jpg.

Ranger: tilts head at biscuit.jpg Woof?

You: Biscuit is a corgi. Customer photo. A very important blob.

Files or Blob? The classic mix-up

Both hold files, so which one? Azure Files behaves like a network drive: many computers can mount the same share at the same time and open files by path, so older apps that expect a file server just work. Blob Storage is object storage that apps and browsers reach through a URL or an application programming interface (API), built to scale to enormous amounts of data. Rule of thumb: people and older apps that map a drive want Files; websites and apps storing millions of objects want Blob.

Exam tip Azure Files shares can be mounted at the same time by Windows, Linux and macOS machines. 'Lift and shift a file share' on the exam means Azure Files.

Standard or Premium

When you create a storage account you choose its performance. Standard is the default, cost-effective choice for most data. Premium uses solid-state drives (SSDs) for consistently low latency and high transaction rates, and costs more per gigabyte. You also choose the account's redundancy (how many copies, and where; that's tomorrow's topic) and, for blobs, a default access tier.

Sam: Premium for the photo app's busy database files, Standard for everything else?

Dev: Now you're thinking like the bill.

The account types

Microsoft recommends four storage account types:

• Standard general-purpose v2: blobs (including Azure Data Lake Storage), files, queues and tables. Recommended for most scenarios
• Premium block blobs: block and append blobs that need high transaction rates, small objects or low latency
• Premium file shares: Azure Files only, for enterprise or high-performance apps; supports SMB and NFS shares
• Premium page blobs: page blobs only

Exam tip Default answer: Standard general-purpose v2. It's also the only type that offers every redundancy option; Premium types offer only locally redundant (LRS) and zone-redundant (ZRS) storage.

Names and endpoints

A storage account name must be unique across all of Azure, 3 to 24 characters long, using only lowercase letters and numbers. The name becomes part of each service's endpoint, the address apps use to reach it. That's why Pawsome_Pets would be rejected, and why someone else may already own 'pets'.

pawsomepetsdata.blob.core.windows.net
pawsomepetsdata.file.core.windows.net
pawsomepetsdata.queue.core.windows.net
pawsomepetsdata.table.core.windows.net

Sam: I tried to name a test account 'kibble'. Taken. 'kibbleco'? Taken. I'm now kibbleco7x2.

Four access tiers

Blob data can sit in an access tier that matches how often it's used. The trade-off: the cooler the tier, the cheaper it is to store data, but the more each read costs.

• Hot: data used often. Highest storage cost, lowest access cost
• Cool: data used now and then; keep it at least 30 days
• Cold: data rarely used that still needs fast reads; keep it at least 90 days
• Archive: data rarely used that can wait hours to read; keep it at least 180 days. Lowest storage cost, highest cost to read

Online and offline

Hot, cool and cold are online tiers: reads take milliseconds. Archive is offline: you can't read or change an archived blob until you rehydrate it to an online tier, which can take up to 15 hours. Moving or deleting a blob before its tier's minimum (30 days cool, 90 cold, 180 archive) adds an early deletion charge for the remaining days.

Exam tip Tiers apply to block blobs. An account's default tier can be hot, cool or cold, never archive; archive is set on individual blobs.

Let rules move the data

Nobody wants to retier two million photos by hand. Lifecycle management policies move blobs to cooler tiers, or delete them, based on rules such as 'not modified for 30 days'. Dev's plan for Pawsome: new photos start hot, move to cool after 30 days, and go to archive after a year; grooming records are archived for their seven-year retention.

Dev: Set the rule once and the old photos walk themselves to the back shelf.

Ranger: trots to the back shelf to check