Other topics
Distributed System communicationCAP
Consistency: Consistency means that all clients see the same data at the same time, regardless of which node they connect to.
Availability: Availability means that any client requesting data receives a response, even if some of the nodes are down.
Partition Tolerance: A partition indicates a communication break between two nodes. Partition tolerance means that the system continues to operate despite network partitions.
Availability
Availability Percentages versus Service Downtime
Availability % | Downtime per Year | Downtime per Month | Downtime per Week |
90% (1 nine) | 36.5 days | 72 hours | 16.8 hours |
99% (2 nines) | 3.65 days | 7.20 hours | 1.68 hours |
99.5% (2.5 nines) | 1.83 days | 3.60 hours | 50.4 minutes |
99.9% (3 nines) | 8.76 hours | 43.8 minutes | 10.1 minutes |
99.99% (4 nines) | 52.56 minutes | 4.32 minutes | 1.01 minutes |
99.999% (5 nines) | 5.26 minutes | 25.9 seconds | 6.05 seconds |
99.9999% (6 nines) | 31.5 seconds | 2.59 seconds | 0.605 seconds |
99.99999% (7 nines) | 3.15 seconds | 0.259 seconds | 0.0605 seconds |
Reliability
Mean time between failures (MTBF) and mean time to repair (MTTR)
Scalability
the ability of a system to handle an increasing amount of workload without compromising performance.
Dimension | Definition | Example |
Size scalability | Ability to add resources to handle more workload | Adding more CPUs to handle more requests |
Administrative scalability | Capacity for multiple users to share a single distributed system | Multiple companies sharing a cloud-based system |
Geographical scalability | Ability to cater to a broad geographical region | A search engine serving users in multiple countries |
Approach | Definition | Example |
Vertical scalability | Scaling up by providing additional capabilities to an existing device | Adding more RAM to a server |
Horizontal scalability | Scaling out by increasing the number of machines in the network | Adding more nodes to a distributed system |
Maintainability
Aspect | Definition | Importance |
Operability | Ease of ensuring smooth system operations under normal circumstances and achieving normal conditions under a fault | Important for maintaining system availability and reducing downtime |
Lucidity | Simplicity of the code base, making it easy to understand and maintain | Important for reducing maintenance time and costs |
Modifiability | Capability of the system to integrate modified, new, and unforeseen features without difficulty | Important for adapting to changing requirements and improving system functionality |
Metric | Definition | Formula | Goal |
Maintainability (M) | Probability that the service will restore its functions within a specified time of fault occurrence | M = probability of restoring the component to its fully active form within a specified time | High M value |
Mean Time To Repair (MTTR) | Average amount of time required to repair and restore a failed component | MTTR = total maintenance time / total number of repairs | Low MTTR value |
Fault tolerance
Technique | Definition | Trade-offs |
Replication | Replicating both services and data to swap out failed nodes or data stores with healthy ones | Consistency vs. availability trade-off, synchronous vs. asynchronous updates |
Checkpointing | Saving the system's state in stable storage when the system state is consistent | Consistency vs. availability trade-off, synchronous vs. inconsistent updates |
Types of Data Center Servers
Server Type | Purpose | Example Resources |
Web Servers | Handle API calls from clients behind the load-balancer (mostly serve static content) | Medium memory and storage resources, good computational resources |
Application Servers | Run core application software and business logic (servers primarily provide dynamic content) | Extensive computational and storage resources, volatile and non-volatile storage, up to 256 GB RAM and 6.5 TB storage |
Storage Servers | Store and manage structured and non-structured data | Structured (SQL) and non-structured (NoSQL) data management systems, storage capacity up to 120 TB, exabytes of storage, 32 GB RAM |
Load balancing
Load Balancing Technique | Local Load Balancing | Global Load Balancing |
Definition | Balancing within a data center. | Balancing traffic across multiple geographical regions. |
Purpose | Improving efficiency and better resource utilization within a data center. | Distributing traffic intelligently across multiple geographical regions. |
Focus | Within a data center. | Across multiple geographical regions. |
Technology used | Reverse proxy. | Load Balancing as a Service (LBaaS). |
Installation location | Within the data center. | Can be installed on-premises or obtained through LBaaS. |
Load balancing technique | Divides incoming requests among the pool of available servers. | Uses techniques such as DNS and round-robin to perform load balancing. |
Limitations | Limited control over the client's behavior, smaller packet size (512 bytes), clients can't determine the closest address to establish a connection with. | Can suffer from uneven load distribution on end-servers, keeping on distributing the IP address of the crashed servers until the TTL of the cached entries expires. |
Use of ADCs | Used as an additional layer of load balancing. | ADCs can implement GSLB. |
DataBase
Feature | Relational Database | Non-Relational (NoSQL) Database |
Data structure | Organized in one or more tables/relations | Can be structured, semi-structured, or unstructured data |
Query language | Structured Query Language (SQL) | Various languages depending on the type of NoSQL database |
Schema | Follows strict schema and requires data to conform to it | Dynamic schema allows for flexible data |
Scalability | Vertical scaling. Or horizontally scale by separating table | Can scale horizontally by adding more nodes |
ACID properties | Provides full ACID compliance | Often sacrifices some level of consistency for availability and partition tolerance |
Use case | Best for structured data with complex relationships and strict integrity constraints | Best for unstructured or semi-structured data and high scalability needs |
Type of NoSQL Database | Description |
Key-value Database | Stores data as key-value pairs using hash tables, with the key serving as a unique or primary key, and values being anything from simple scalar values to complex objects. Efficient for session-oriented applications, such as web applications. Example databases include Amazon DynamoDB, Redis, and Memcached DB. |
Document Database | Designed to store and retrieve documents in formats like XML, JSON, BSON, etc. Documents are composed of a hierarchical tree data structure that can include maps, collections, and scalar values. Suitable for unstructured catalog data and content management applications. Example databases include MongoDB and Google Cloud Firestore. |
Graph Database | Uses the graph data structure to store data, where nodes represent entities, and edges show relationships between entities. Allows storing data once and interpreting it differently based on relationships. Suitable for social applications, data regulation and privacy, machine learning research, and financial services-based applications. Example databases include Neo4J, OrientDB, and InfiniteGraph. |
Columnar Database | Stores data in columns instead of rows, enabling access to all entries in the database column quickly and efficiently. Suitable for large numbers of aggregation and data analytics queries. Example databases include Cassandra, HBase, Hypertable, and Amazon SimpleDB. |
Terms:
- Cloud computing is running applications on computing resources managed by cloud providers. When using cloud computing, we do not have to purchase or manage hardware ourselves.
- Serverless computing builds on the convenience of cloud computing with even more automation. It enables developers to build and run applications without having to provision cloud servers. The serverless provider handles the infrastructure and automatically scales the computing resources up or down as needed. This provides a great developer experience since developers can focus on the application code itself, without having to worry about scaling.
