Lec 2: Cloud Infrastructure

Lecture 2: Cloud Infrastructure

Cloud Computing: formal definition

Cloud computing is a model for enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable computing resources that can be rapidly provisioned and released with minimal management effor or service provider interaction.

Focus on computing as a utility: Computing is outsourced to a third party or internal organization

 

Essential characteristics:

• On-demand self service
• Broad network access
• Resource pooling
• Rapid elasticity = automatic scalability
• Measured service

Problems of on-premise

On-Premise: We build our own servere

It's difficult to dimension because load can vary considerably

• provisining for the peak load
  • waste of resources when the over provisioning -> inefficient
• provisioning below the peak
  • still over provisioning occurs but less
  • However, cannot service when under provisioning -> dissatisfied customers will leave

Problems

• expensive expecially for hardware
• need expertise and maintenance
• scaling up and down is difficult

Benefits of Cloud

• reduced costs
• usaged-based pricing -> no minimum or up-front fee, helpful when apps have variable utilization
• increased scalability
  • cloud can instantly and dynamically allocate IT resources to consumers on-demand
• economies of scale -> cheaper usage fee for cloud customers
• increased reliability and availability
  • availability -> accessible for longer periods of time
  • reliability -> able to better avoid and recover from exception conditions
  • Cloud providers offer resilient IT resources for which they are able to guarantee high levels of availability (e.g. service level agreement: SLA)

Risk and challenges of cloud

• increased security vulnerabilities
  • cloud provider have privileged access to consumer's data
  • malicious cloud consumers
• reduced operation governance control
  • consumers are usually allowed a level of governance control that is lower than that over on-premise IT resources
• limited portability btw cloud providers
  • challenging to move from one cloud provider to another
• multi-regional compliance and legal issues
  • different regions have different regulations

Cloud infrastructure

Server

• servers are computers that provide services to clients

Blade Server

• compact server
• to conserve the amount of floor space

Racks

• Blade servers are placed in racks
• a single rack can hold up to 42 1U blade servers

Heterogeneous Servers

• all servers are not homogeneous == servers can have different internal hardware configurations
• configurations are optimized for major app classes

Networking

• cloud resources are delivered to users over the Internet
• physical cables, switches, routers, wireless access points ...

Networking Topologies

• Bus: single central cable connected a number of devices
• Tree -> datacenter usually use tree topology : root node, hierarchy
• Ring: closed loop
• Star: hub

 

Storage

Storage Array

• Storage arrays are configured in such a way that they can present storage to multiple servers, typically over a dedicated network.
• storage arrays use multiple drives in a collection capable of storing a huge amount of data, managed by a central management system.

Cluster

• many racks form a cluster
• in rack, there are many nodes/blades
• network switch: connects nodes with each otehr and with other racks
• storage devices

Datacenter

• large server and storage farms connected together
• Giant warehouse filled with racks of servers, network switches, storage arrays
• have cooling system, redundant power, fire protection, physical security, monitoring systems
• often globally distributed because of the need to be close to users, cheaper resources, protection against failure
• large clusters need massive cooling -> datacenters consume a lot of energy