Getting Started with Cloud Performance Tuning - by Tanisha and Prachi

Figure 1

What is Cloud Performance Tuning?

Cloud Performance Tuning is a process that focuses on improving the systems response with time without any need of changing or upgrading the system's configuration.In computer systems,the motivation for such activities is called Performance Problem while which can be real whereas most of the systems will respond to increase the loads system with some degree of decreasing its performance. The capacity of systems to handle higher loads is called scalability and device modification which can handle a higher load is called Performance Tuning.

Steps that are systematic to tuning:

1. Assess the problem and set numerical values that qualify as an appropriate.
2. Before modifying of the system measure the performance.
3. Identify the parts of the system which is the key to performance improvements.This is called bottleneck.
4. To remove the bottleneck, you should modify the parts of the system.
5. Measure the system adjustment performance after its modification.
6. And if the modification of the system makes its performance better, adopt it else put it back the way it was in the system.

This is an instance of the measure-evaluate-improve-learn cycle for the quality assurance which usually means the way of preventing mistakes.A performance problem may be identified by slow unresponsive systems which usually occurs because of the high system loading that means measure of the amount of computational work that a system performs, causing some part of the system to reach a limit to response which means bottleneck.

A bottleneck is the ability of the capacity of an application or the computer system which usually limits by a single component like a neck of bottle slowing down the overall water flow.

The bottleneck has the lowest throughput of all the transaction path

Figure 2

How can we improve cloud performance tuning?

1. Check the Application: Check first about and demand of the application. Low output in cloud workload is also blamed on the insufficient resources but that matters is a poorly designed,poorly managed, poorly programmed or poorly deployed applications which a simple code and design changes to good to solve most of its performances issues.
2. Check Intraclouds Network latency: If we assume that intra cloud network bandwidth will exceed any workload requirements which if not always the case.Many times when the workload is decoupled from the data source. performance issue is the bandwidth which is the measurement of the amount of information that a particular computer network or the internet connection can be send per unit time i.e., bits per second between the machine instances hosting the workloads and the data source, be it intra clouds or inter clouds.

Also it is complex to check the bandwidth intracloud because the open internet usually have burst speeds.

      3. Check the database: Some of the applications that are mostly of database performance issues, or cloud based or not, which  has been come from the poorly designed database usually not a slow one and to improve database performances or  tuning will all depend largely on the database while most  consider index utilization and caching schemes in increasing the workloads.

CLOUD LOAD BALANCING AND ITS IMPORTANCE

Load balancing is a key characteristic of the cloud computing model. Applications often need to scale up and down to meet varying load condition. It is a highly desirable feature of iaas clouds.

It also reduces costs associated with document management systems and thus maximizes availability of the resources.

Techniques of the Load Balancing:

Scheduling Algorithm: It allows users to set conditions for when they want there applications to scale up and down based on application-specific matrices such as transactions per second, number of simultaneous user, request latency.

Load Balancing Policies: It specifies the unique and special property (USP) of requests and computing nodes which is implemented in a dis-centralized manner with low overhead called Workload and Client Aware Policy(WCAP) and with the help of USP, a scheduled that can be decided the most suitable node to be compete in time.

A Comparative Study of Algorithms: Application performance should be guaranteed. By using powerful resources and infrastructure that increases performance and optimization and then deliver the complete services to cloud user.

How does Cloud Performace Tuning Work?

Points that need to be take:-

Figure 3

Top Performance(Load) Testing tools in 2020

1. Webload
2. StesStimulus
3. LoadNinja
4. LoadView
5. NeoLoad
6. RedLine13

And many more ....

Benefits of Cloud Testing :

It supports :

1. Testing Scalability
2. Geographical Testing
3. Testing Production Apps and Systems
4. And Reduces the overall Costs.

References:

Figure

1: https://www.cloudibn.com//

2: https://www.ibm.com//

3: https://www.testingservices.com//