You design data engineering solutions for a company that has locations around the world. You plan to deploy a large set of data to Azure Cosmos DB.
The data must be accessible from all company locations.
You need to recommend a strategy for deploying the data that minimizes latency for data read operations and minimizes costs.
What should you recommend?
A
With Azure Cosmos DB, you can add or remove the regions associated with your account at any time.
Multi-region accounts configured with multiple-write regions will be highly available for both writes and reads. Regional failovers are instantaneous and don't require any changes from the application.
Reference:
https://docs.microsoft.com/en-us/azure/cosmos-db/high-availability
You need replicate data only for reading and minimizes costs. So I think the correct answer is B. Use a single Azure Cosmos DB account Configure data replication. Reference: https://docs.microsoft.com/en-us/learn/modules/distribute-data-globally-with-cosmos-db/2-global-distribution
The highest voted answer is wrong. That kind of things doesn’t help. The answer given is correct.
Correct answer: B. Requirement is *minimized read latency at minimal cost*. Both *Data replication* and *Multi-region writes* reduce the read latency - because both options increase the number of *read regions*. Larger number of read regions => less latency. - *Data replication* allows you to have one write region and *multiple read regions*. - *Multi-region writes* is an additional option under *Data replication* that allows you to have multiple *write regions*. It also increases the number of *read regions*, of course. But multi-region writes offer more than we need here, since we do not have a requirement for faster write requests. *Multi-region writes* is a more expensive option because write requests are more expensive than read requests: - Single-region write account distributed across N regions: 0.008$ - Multi-region write account with N regions: 0.016$
Pricing ref: https://azure.microsoft.com/en-us/pricing/details/cosmos-db/
Pricing ref: https://azure.microsoft.com/en-us/pricing/details/cosmos-db/
another proof here : https://docs.microsoft.com/en-us/azure/cosmos-db/high-availability#multi-region-accounts-with-a-single-write-region-write-region-outage
the given answer is correct. Enable multi-region writes
Why ? there are not write requirements but there are reading requirements so i think it should be B
but part of one of the requirements is "and minimizes costs." would data replication be cheapest way to accomplish? I think given answer is correct.
Multi write will cost 3 times more than single write, multi read in this scenario.
Multi write will cost 3 times more than single write, multi read in this scenario.
There is no option for Data Replication. You can only set GeoRedundancy o Enable multi-region writes
it's this option on Azure portal > Cosmos DB Account > "Replicate data globally"
it's this option on Azure portal > Cosmos DB Account > "Replicate data globally"
but part of one of the requirements is "and minimizes costs." would data replication be cheapest way to accomplish? I think given answer is correct.
Multi write will cost 3 times more than single write, multi read in this scenario.
Multi write will cost 3 times more than single write, multi read in this scenario.
There is no option for Data Replication. You can only set GeoRedundancy o Enable multi-region writes
it's this option on Azure portal > Cosmos DB Account > "Replicate data globally"
it's this option on Azure portal > Cosmos DB Account > "Replicate data globally"
Believe the correct answer should be B.
So far, C is wrong because it will only replicate to 1 paired region. A seems to be overkill - there is no write redundancy requirement, and if current write region fail, Cosmos will auto failover to one of the read regions and will make it write enable. So B looks like a right answer to me.
The answer is correct. The requirement is "minimizes latency for data read". The table (see link below) shows "low" read latency for Multi-region writes while "Cross region" for the rest. https://docs.microsoft.com/en-us/azure/cosmos-db/high-availability#availability-zone-support
With B there z latency
Assuming that "the company has locations around the world" means there is multi-region writes required for the solution, the given answer is correct. I think the "minimizes latency for data read" requirement is the confusing part. But multi-region writes would minimize the read latency. On the other hand, "and minimizes costs" part seems not to be covered by this solution. All in all, I think A is a valid answer and in real-world circumstances, it should be used for international companies.
once data is replicated, the read consistency becomes 99.999 i.e. reads can happen from any region, for writes to offer 99.999 one must enable multi-master model, by enabling multi-region writes for other replicated regions. So B is appropriate here.
Why not D. Use a single Azure Cosmos DB account. Enable geo-redundancy.? As we need to minimize cost and we only need read not write
Because you will need RA-RGS not just RGS.
Geo-redundancy is more related to failover mechanishms. what you need is a data replication technique here. You need to have single write region and multiple read regions to minimize the read latencies across the globe
B. Use a single Azure Cosmos DB account Configure data replication.
I think the correct answer is B. Reference: Whizlabs Course!
The correct answer is : B : Use a single Azure Cosmos DB account Configure data replication.
With multi-region writes, you can set session consistency (which read your own writes). With this you can minimize latency to read/write around the world and also required less RU's compare to Strong consistency so it minimize the cost as well. Another thing is if the application is being use around the world then read/writes happening around the world. In that scenarios, you need multi-region writes to minimize the latency and good partition key to for performance and reduce the costs for read/write operations.
From the reference below it appears A is the correct answer. To ensure high write and read availability, configure your Azure Cosmos account to span at least two regions with multiple-write regions. This configuration will provide the highest availability, lowest latency, and best scalability for both reads and writes backed by SLAs. To learn more, see how to configure your Azure Cosmos account with multiple write-regions. https://docs.microsoft.com/en-us/azure/cosmos-db/high-availability#building-highly-available-applications
but what about the costs?
Single-region accounts may lose availability following a regional outage. It's always recommended to set up at least two regions (preferably, at least two write regions) with your Cosmos account to ensure high availability at all times. https://docs.microsoft.com/en-us/azure/cosmos-db/high-availability I think given answer is right.
cost of replicating to a region is the same as the original region, so replicating to 3 additional regions, would cost approximately four times the original non-replicated database
B is the correct answer. Requirement is Read availability at minimal cost
https://docs.microsoft.com/en-us/azure/cosmos-db/tutorial-global-distribution-sql-api?tabs=dotnetv2%2Capi-async Answer is B
I think answer is correct. If you configure multiple region writes, updating of data might be delayed based on consistency level configured but reads will be faster for already replicated data.
"for already replicated data" yes there is another answer for it