Welcome back. This is part two of our discussion on storage accounts. Let's take a look at our learning objectives.
My objectives include discussing performance, tears, access, tears and storage replication when creating storage count. Specifically general purpose storage accounts. We do have some options for performance.
Our first option is standard performance. These air backed by magnetic hard disk drives and are the lowest cost per gigabyte available. Standard performance is for applications that require bulk storage or where data is accessed less frequently. The next option is premium performance.
These air, backed by solid state drives or SS D's
premium storage, is a high performance low Leighton see option for intensive input output operations.
Premium performance can only be used for virtual machine disc.
Be EMS using premium storage will have a 99.9% S L A.
In terms of our storage accounts, the general purpose storage account can use standard performance for blobs, files, tables, cues and VM discs. And can Onley use premium for unmanaged virtual machine disc
blob storage accounts can use standard performance on Lee
Now I've mentioned access tears a few times, so let's get an understanding of what that ISS access tears provide different options for accessing data based on the frequency of the usage. Each tier is optimized for a particular pattern of data. By selecting the appropriate here, you can store your data in a cost effective manner.
The first here is hot tear. This is optimized for frequently accessed objects in the storage account.
The second tier is cool. This is for data that is infrequently access and is stored for at least 30 days.
Storing data here may be more cost effective, but accessing the data is more expensive than the hot tear. The last here is archive. This is available on Lee for individual block blobs and cannot be set on the entire storage account.
This is for data that could accept several hours of Leighton see while being retrieved. It is recommended that data remain in the archive tear for at least 100 and 80 days. It is the most cost effective option, but it has the most expensive access costs.
Luckily, you can switch access tears at any time. If there's a change in the usage pattern of your data,
Our third characteristic on storage accounts is replication, and we have several options to choose from when creating our storage accounts.
The first is locally redundant storage, or LRS. This is the lowest cost of our replication options, and the data stored in the storage account is replicated to a single storage scale unit in an azure region. The next type is own redundant storage or Z R s.
This has a higher availability. As the data is synchronously replicated across three availability zones in an Asher region.
Our next replication option is geo Redundant Storage, or GRS. GRS takes a cross region replication option by replicating our data to another azure region several 100 miles away.
This will protect you against a region wide outage in an azure data center. A variation on GRS is the reed access Geo Redundant Storage or R A G. R s. This is the same as GRS.
But before Microsoft initiates a fail over to the other set of data and another azure region,
you will have read access on Lee to that replica. Now, understand these options maybe a little bit confusing and hard to understand just by looking at this table. So let's jump over to our white board to get a better understanding of these replication options.
Let's say we have a storage account created in the East US region
and inside of this region. There are multiple availability zones
for locally redundant storage, or LRS. There are three copies of the data in the storage account. In one availability zone,
the data is replicated synchronously inside the zone. Riding to the storage account is considered complete and finalized when all replicas have committed the changes.
Now let's take this a step further by talking about Zone redundant storage and zero RS. Instead of having copies of the data in one availability zone,
a second and third copy the data is made to two different zones inside the same region.
This protects against own failures. CRS also uses synchronous replication, ensuring each copy across the zones are the same.
Next is Geo redundant storage. For this, we need another region several 100 miles away. Like the West US
for GRS, we go back to having our data and one availability zone in the East US.
We also have three copies in one available to zone in the West US. It's like having two lRS deployments and two different regions.
This data is copied a synchronously, meaning there is a delay after the data is committed in one region before it is replicated to the secondary region in the event of a region. Why disaster? Any changes that have not replicated to the other region may be lost.
The data in the second region is not available for reader right access until Microsoft initiates a fail over.
This is where Reed Access GRS comes into play
with this option. Data in the second region is available through read on Lee Access Before Microsoft initiates a fail over
That's it for this episode, hopefully have a better understanding of some of the options. We have one creating storage accounts
before we jump to the next step. So let's take a quick knowledge check which performance tear do blob storage account support.
You said Standard performance. You are correct.
And another question
Which replication option involves the second region?
You said geo redundant storage or read access Geo Overdone the storage. You are right. Both of these replication options will replicate your data to a second azure region several 100 miles away.
Moving to the next episode. We're gonna take some of these concept we've been learning and actually see what they look like inside the azure portal when we're creating and configuring storage accounts, See you in the next episode.