• When there are huge surges in demand, it is preferable to user serverless technologies that automatically scale on demand. In this case, the key concern is the data ingestion. Pub/Sub is a serverless system that can expand to absorb such demand.
• Remaining options could all potentially over-estimate or under-estimate the demand, resulting in either a loss of money or unmet demand.

• Data Fusion, a codeless environment for creating data pipelines using a drag-and-drop interface. Data Fusion actually uses Dataproc as its Compute Engine, so everything we have mentioned earlier applies also to the case of Data Fusion. If your team prefers to create data pipelines without having to write any code, Data Fusion is the right tool.
• Dataflow – you have to write code. Either Java or Python, as of this writing.
• Dataproc – is used to run Hadoop/Spark workloads. Requires coding.
• Datastream – this is not used for data processing. Used to replicate and synchronize data.

References:

1. https://cloud.google.com/blog/products/data-analytics/building-the-data-engineering-driven-organization
2. https://cloud.google.com/data-fusion

• Option A is correct. Cloud logging helps in real-time management and analysis of logs.
• Option B is incorrect. Cloud Monitoring helps monitor the health, availability and performance of infrastructure and the configured application by producing metrics of help.
• Option C is incorrect. Cloud Trace is an offering from Google Cloud that helps find production performance bottlenecks by collecting applications latency data.
• Option D is incorrect. Cloud Debugger helps investigate the behavior of code in production.

References:

Logging: https://cloud.google.com/logging

Monitoing: https://cloud.google.com/monitoring

Trace: https://cloud.google.com/trace

Debugger: https://cloud.google.com/debugger

Reference: https://cloud.google.com/billing/docs/how-to/budgets

• 
  
• Since we are experimenting, we don‘t have a good estimate on how much processing will be done. With the on-demand pricing model, you are charged for the number of bytes processed by each query. The per-byte-scanned billing model can be advantageous in that you pay precisely for what you use, with the amount of scanned data as a proxy for compute consumption.
• 
  
• BigQuery Monthly and Annual reservations can be used for cost predictability and to provide guaranteed capacity for workloads with SLAs.
• 
  
• BigQuery Flex Slots for cyclical workloads that require extra capacity, or for workloads that need to process a lot of data in a short time, and so would be less expensive to run using reserved slots for a short time.
• 
  
• 1 TB per month might not cover your analytics needs, except for the smallest, temporary projects. It‘s better to plan to get better pricing.

• A partitioned table is a special table that is divided into segments, called partitions, that make it easier to manage and query your data. By dividing a large table into smaller partitions, you can improve query performance, and you can control costs by reducing the number of bytes read by a query.
• 
  
• Splitting the table is a possibility but now you have to maintain multiple tables. You will also have to rewrite the upstream data processing tool to deliver the data to the corresponding table.
• 
  
• Optimizing the query could have some benefits. But if the query is still processing all the data, there will be limits to the optimization you can achieve.
• 
  
• Cloud Spanner is built for transaction processing rather than analytics processing.

• Compute Engine provides graphics processing units (GPUs) that you can add to your virtual machines (VMs). You can use these GPUs to accelerate specific workloads on your VMs such as machine learning and data processing.
• Even when using compute optimized machines, they still use CPUs. Those are great for general computing tasks. GPUs are optimized for number crunching.
• Most of this compute is expected to be in-memory. Disk usually doesn‘t play much of a role.
• Cloud Functions support general purpose applications. They too will be running CPUs. There is no way to change that to GPUs.

References:

1. https://cloud.google.com/compute/docs/gpus
2. https://cloud.google.com/compute/docs/gpus/create-vm-with-gpus

https://cloud.google.com/support/ <https://cloud.google.com/support/

• Option A is incorrect. Snapshot can be created from disks when attached to a running VM. Hence, this statement is incorrect.
• Option B is correct. Snapshot data integrity is ensured by automatic checksum.
• Option C is incorrect. Snapshot data integrity is ensured by automatic checksum. Hence, the state of the VM is not a concern.
• Option D is incorrect. Snapshot data integrity is ensured by automatic checksum. Data is incrementally backed up from disks.

Reference: https://cloud.google.com/compute/docs/disks/snapshots

Compute Engine is Google Cloud's main compute service, providing scalable, on-demand virtual machines (VMs) that can be used to run a wide variety of applications and services. Compute Engine provides a high level of control and customization, allowing users to choose the operating system, software, and other configuration options that best meet their needs. Additionally, Compute Engine integrates with other Google Cloud services, such as Google Cloud Storage and Bigtable, making it easy to build complex, multi-tier applications that can scale as needed.