Enterprise Computing Flashcards
What is the waterfall model?
A linear, sequential software development approach with distinct phases: Exploration -> Development -> Operations
What is the interactive/incremental model?
Software Development Lifecycle involving iterations, each delivering small, functional increments.
What is the Learn phase of the Learn-Measure-Build cycle?
An enterprise comes up with a hypothesis about the marketplace and decided what empirical data would validate this hypothesis.
What is the Measure phase of the Learn-Measure-Build cycle?
An enterprise tests its hypothesis by collecting the empirical data.
What is the Build phase of the Learn-Measure-Build cycle?
An enterprise creates a Minimal Viable Product if the empirical data looks good.
What is a Technology Pivot?
When it becomes clear that the product could deliver its value more efficiently using different technology.
What is a Zoom-In Pivot?
When a product feature becomes a product.
What is a Zoom-Out Pivot?
When a product becomes a product feature.
What is a Customer Segment Pivot?
Changing the target audience while keeping the product roughly the same.
What is a Customer Need Pivot?
Changing the product to address a different need for your original customer group.
What is a Concierge MVP?
When a person replicates how the product would work, and the customer knows someone is replicating it.
What is a Wizard of Oz MVP
When a person replicates how the product would work, however the customer does not now a person is replicating it.
What is a Landing Page MVP?
A page where potential customers can find out about the product idea, and possibly pledge money.
What is a Video MVP?
A video that shows how the product might work, and asks customers to sign up for it.
What sort of experiment is a startup?
An experiment where you have a hypothesis you are trying to test.
What is the biggest waste that product development faces today?
Building things very efficiently that no one wants.
What is the universal constant of all successful startups?
The pivot.
Is agile development right for a startup?
No as with agile, the customer is knows and the specification is not, however with a startup neither the customer or specification is known.
What is Validated Learning?
Learning quantitatively how to build a sustainable business.
What needs to be in the first version of a product?
Only what is necessary to learn whether the plan is correct or not.
What should the heuristic be for any kind of startup advice?
Does it minimise total time through the build-measure-learn loop?
What are actionable metrics?
Metrics about per customer behaviours that can be measured.
What is a monolithic system?
A single program run by a single process formed from a collection of modules that communicate by procedure calls.
Why is it easy to develop a monolith?
Because all the code is in one language in one place.
Why is it easy to test a monolith?
Because the system can be a single executable against which a suite of tests can be run automatically.
How do you scale a monolith?
Vertical scaling, which means making the machine more powerful.
What database will a monolith use?
A centralised database, meaning that the accuracy, completeness, and consistency of data is maintained.
What are the four ACID properties?
Atomicity, Consistency, Isolation, Durability
What is Atomicity?
Transactions either succeed completely, or fail completely, even in the presence of system failures.
What is Consistency?
Transactions begin with the database in one consistent state, and end with the database in another consistent state.
What is Isolation?
The effect of performing transactions concurrently is the same as performing them sequentially.
What is Durability?
Once a transaction has succeeded, its effect persist, even in the presence of system failures.
What is the Plan to throw one away?
The first version of a system should always be viewed as a prototype to be discarded once a proper understanding of what is wanted has been obtained.
Why is Gall’s Law?
A complex system that works is invariably found to have evolved from a simple system that worked. A complex system designed from scratch never works and cannot be patched up to make it work. You have to start over with a working simple system.
What is the You Aren’t Gonna Need It observation?
Functionality thought to be needed in future should not be be built now as you are not going to need it. This is because it incurs a build cost, and a delay cost for other functionality.
What is Conway’s Law?
Any organisation that designs a system will inevitably produce a
design whose structure is a copy of the organisation’s
communication structure.
What is the observation about eating your own dogfood?
The creators of a system should use it themselves to demonstrate its usefulness and usability, and improve its quality through bug discovery.
What lesson can be learned from the evolution of large companies?
No one starts with micro services, however past a certain scale everyone ends up with it. However most don’t reach that scale.
What is the correct architecture for the start of an enterprise?
A monolith that is just enough to meet near-term customer needs.
What does just enough architecture look like?
Simple technology that is familiar to the dev team, often rapid prototyping framework.
Why should one buy not build?
Doing so is faster, cheaper, and better than people can do themselves.
What architecture is appropriate for the scaling phase of an enterprise?
A microservices architecture where teams can design, develop, deploy and operate their services independently.
What should incremental changes look like?
As small as possible, to maintain backward/forward compatibility of data and interfaces.
What is the system of record?
The single service that owns any piece of data - every other copy is a read only, non-authorititive cached one.
What architecture is appropriate for the Optimising phase of an enterprise?
One that is more stable because it makes only sustainable, incremental improvements in functionality and efficiency.
What is a Microservices system?
A system that consists of multiple programs that run as multiple processes and communicate by sending messages over a network.
What is REST?
Representational State Transfer, a form of HTTP.
What are the four HTTP methods used in REST?
GET, PUT, DELETE, POST
What are the four resource archetypes in REST?
Document, Controller, Collection, Store
What is the “document” resource archetype in REST?
A file-like resource where: GET reads the resource, PUT updates it, DELETE removes it.
What is a “controller” resource in REST?
A resource that triggers an action or task when a POST request is made.
What is a “collection” resource in REST?
A directory-like resource where: GET lists items in the directory, and POST creates a new item with an invented name.
What is a “store” resource in REST?
A directory-like resource where: GET lists items, and PUT creates or updates an item with a specified name.
What is a distributed database?
A distributed database is a collection of data stored across multiple physical locations, often on different servers or data centres, that appears to users as a single database. However it makes it difficult to ensure that the accuracy, completeness, and consistency of data is maintained.
What are the two ways to manage a distributed transaction?
Two-phase commit or sagas.
How does a Two-phase commit work?
A coordinator transaction asks participant transactions to vote on committing a change. Each participant locks its data until a decision is made. If all vote to commit, the coordinator instructs all to commit; if any vote to abort or time out, the coordinator instructs all to abort.
How does Sagas work?
In a Saga, a coordinator calls each participant in sequence to either commit or abort. Each participant only locks its data until it decides. If all participants commit, the transaction is successful.
If any participant aborts, the saga stops immediately and compensating transactions are executed to undo the committed actions of earlier participants.
This ensures eventual consistency without long-held locks.
What are the pros and cons of a Two-phase commit?
Pros:
- Ensures strong consistency
- Automatic rollback if coordination fails
Cons:
- Locks can be held for a long time
- Single point of failure in the coordinator
What are the pros and cons of sagas?
Pros:
- No long held lock
- Scales better for long running processes
- More fault tolerant
Cons:
- Only guarantees eventual consistency
- Requires writing compensating logic for rollbacks
How do you scale a system of Microservices?
Horizontal scaling, which involves adding more machines which can run one or more microservices instances
What is the Strangler Fig Design Pattern?
A method for migrating from a monolithic MVP to a microservices one. Initially the modules of the monolith are put behind a facade, which serves as a proxy that manages communication between them. Modules are then replaced by microservices one at a time, updating the facade each time.
What is the second system effect, and how may it apply to a team migrating to microservices??
After building a minimal first system, developers often try to include every feature and idea they previously left out — leading to excessive complexity and poor design. A team migrating to microservices should be wary of losing focus and creating something not required.
How does Jeff Bezos’ “two-pizza team” rule apply to migrating from a monolithic MVP to microservices?
Bezos’ rule suggests that teams should be small enough to be fed with two pizzas. Each microservice should be worked on by a small, independent team, under the direction of a leader who gives the service undivided attention.
How does the “Big Ball of Mud” concept by Brian Foote and Joseph Yoder apply to teams migrating from a monolithic MVP to microservices?
The observation about the Big Ball
of Mud is that over time the design of a system may decay to a big mess, and therefore a team migrating to microservices may choose to start the migration when they can no longer work with the big ball of mud.
What is technical debt, and how may it apply to a team migrating to microservices?
Technical debt is when code is written in a quick and dirty function which works short term, but is costly in the long term. When moving to microservices they may choose to pay the technical debt.
What is the CAP theorem, and how may it apply to teams migrating to microservices?
It is that a team migrating to microservices must choose any two of:
Consistency - All clients see the same data at the same time, no matter which node they connect to.
Availability - Any client making a request for data gets a response from a node, even it other are down.
Partition-Tolerance - The system must continue to work despite communication breakdowns between nodes.
What are the 9 common characterisitics of microservices?
- Componentization via services
- Organised around business capabilities
- Products not projects
- Smart endpoints and dumb pipes
- Decentralised governance
- Decentralised data mamagement
- Infrastructure automation
- Design for failure
- Evolutionary design
What is a component?
Something that is independently replaceable and independently upgradeable.
Why should one organise around business capabilities?
So that each service team connects right through to service users and that is judged on how it affects business outcomes.
Should endpoints be smart or dumb?
Endpoints should be smart, and pipes should be dumb. This means logic should live in the endpoints, and the communication channels should simply transfer it.
What is the rule for microservice data management?
Every service should be responsible for its own data and persistence.
What do you have to assume in any distributed system?
That things are going to break.
How big is a microservice?
Something that can be developed by a team that can be fed with two pizzas.
What must be sorted out before going down the microservices route?
- Rapid provisioning (Ability to quickly spin up environments and infrastructure)
- Basic monitoring (Visibility into service health, failures, and performance)
- Rapid application deployment (Automation for fast, reliable deployments)
- DevOps culture (Collaboration between development and operations with shared responsibility for delivery and uptime)
