Enterprise Computing - Operations Flashcards
(51 cards)
What is a staging environment in cloud computing?
A scaled-down replica of the production environment used to test MVPs before release. The infrastructure that provides this is rented in the cloud.
What are the five key features of cloud computing success?
- Broad network access
- On-demand, self-service
- Measured service
- Rapid Elasticity
- Resource Pooling
What does “Broad Network Access” mean in cloud computing?
That the cloud is available over standard networks, including VPN’s.
What is the “drug-dealer” pricing model in cloud computing?
A model where customers can easily and immediately acquire virtual machines on demand.
How and why is usage monitored in cloud computing?
The provider monitors and measures usage for optimization and billing purposes.
What is rapid elasticity in cloud computing?
The ability to scale resources up or down based on demand.
Define resource pooling in cloud computing.
Assigning virtual machines to physical ones so that several users can use the same physical machine.
What are the two phases of cloud computing?
Serverful and Serverless Computing
What are the models of serverful computing?
Infrastructure-as-a-service: Access to bare servers; Platform-as-a-service: Access to servers with operating systems and tools; Software-as-a-service: access to applications on a subscription basis.
What are the main technologies used for serverful implementation?
Virtual machines and containers.
What is the cost model of serverful computing?
Charges are based on resource allocation, just like renting a car.
What is the role of a hypervisor in VM management?
It maps virtual to physical resources, ensuring fair sharing.
How do containers differ from VMs in resource allocation?
Containers share the host OS, mapped by the OS itself.
How can microservices be deployed in serverful environments?
One microservice per VM or one microservice per container.
What are the models of server less computing?
Backend-as-a-Service: Access to services such as authentication or database storage. Function-as-a-service: Cloud provider runs custom code in response to requests or events.
What does “serverless” actually mean?
Servers exist, but their management is abstracted away by the provider.
What is the serverless cost model?
Pay-as-you-go based on execution time, like hailing a taxi.
How can microservices be implemented in serverless environments?
One microservice to one function instance, or one microservice to multiple function instances.
What are challenges with mapping one microservice to many function instances?
Maintenance problems (keeping track of instances), and performance problems (keeping ‘warm’ instances).
According to Wells, how long did it take to get a server ready for code deployment at the FT in a data centre vs AWS?
120 days in an FT data centre; minutes in an AWS data centre.
According to Wells, should one worry about vendor lock-in?
No; it’s more important to avoid delays from doing everything in-house.
According to Wells, what was the FT’s deployment frequency before and after moving to the cloud?
12 releases per year before; about 30,000 changes per year after.
According to Wells, do you have to choose between speed and stability?
No; moving fast actually helps fix issues quicker and break fewer things.
According to Wells, why should you use a queue?
To decouple producers and consumers, avoiding tight synchronous dependencies.
