IT601 Paper

IT601 Notes · All in One (All eggs in one basket) · • What it means:

Using one main server to handle all services like: o Database o Web o DNS (Domain Name System) o HTTP (Web requests) o Proxy (Internet access manager).

• Key Features:

o Uses high-end hardware because one server does all the work. o Virtual Machines (VMs): These are software-based systems that behave like physical computers. o Containers: Smaller, faster units for running applications.

• Problems with All-in-One Setup:

o Complete failure: If the main server crashes, all services stop working. o Complex maintenance: Hard to manage multiple services on one machine.

Snowflake Architecture

What is Snowflake Architecture?

• Definition: Using separate machines for different services.

o Buy servers as needed (exactly what is required for a task). o Each machine is built to match the specific service, e.g.: ▪ Amount of RAM, disk size, network speed, and space for future upgrades. o The machines work efficiently as they are customized for the job. o Vendors compete to offer the best machine for the specifications.

• Disadvantages:

o You end up with a collection of unique machines. o Managing all these machines (a "fleet") can be challenging.

Challenges with Snowflake Architecture

• Administrative Overhead:

o Each machine might use a different RAID system (used for managing disk storage). o IT staff must learn to configure, repair, and update many different systems.

• Solution: o Standardize hardware and software across all machines. o Use a common RAID product to reduce learning curves and improve efficiency.

Asset Tracking in Snowflake Architecture

• Why is it important?

Managing many unique machines requires a proper inventory system to avoid confusion.

• What should be tracked?

o Technical details like operating system and hardware specs (RAM, CPU, etc.). o Ownership details:

▪ Who owns the machine?

▪ Who to contact when something goes wrong?

o Services being used on the machine.

• Best Practices:

o Use automation to keep information up to date. o If automation isn’t possible, conduct annual reviews to identify outdated data.

Reducing Variations

• Why reduce variations?

o Managing too many different types of machines and systems makes work harder.

• How to reduce variations:

o Pick a default vendor, hardware model, and operating system. o Make it easy for people to order and use these defaults. o Provide automated tools for installing and configuring operating systems. o Create a wiki to share recommended models, vendor contacts, and guidance.

Global Optimization vs. Local Optimization

• Local Optimization:

o Customizing each machine for specific tasks might seem efficient. o Example: Adding a machine from a new vendor could require learning new firmware and processes.

• Global Optimization:

o Standardize machines from one vendor to save time and reduce complexity. o Reduces the need to manage spare parts, support processes, and training for multiple systems.