Balancing The Workload In A Cloud Environment | Part II

Types of Market Segments

For each type of cloud service, the policy limit varies from one market segment to another. The policy can be influenced by type of organization, organization size, market conditions, workload demands seasonal economy, mandates changes, emerging
technologies and frequency of adverse weather conditions.

The number also depends on the data center market segment. For example, the government sector is a heavy user of data centers and has been seeking ways to save costs, hiring on demand services to ensure availability and security of operation in the cloud computing environment.

I have already mentioned some market segments as examples – retail, energy and utilities, financial markets, e- governance services, health care, telecommunications and petrochemicals. Besides these, there are others.

  • Aerospace and defense
  • Automotive
  • Construction
  • Consumer Products
  • Education
  • Electronics
  • Forest & Paper
  • Insurance
  • Biological Sciences
  • Media and entertainment
  • Metals & Mining
  • Tourism and transport
  • Manufacture and assembly
  • Industrial products
  • Biological Sciences
  • Shipbuilding
  • Distribution and wholesale services

Let’s compare and contrast the retail market segment and petrochemical-based considerations for policy limits. When the system detects each market segment demands of the workload that exceeds the threshold level, the system quickly creates
additional instances of resource demands to balance the workload dynamically. As the demand of the workload falls below the threshold level, the instances of resources that were allocated are released.

The segment of the retail market consists of small and large companies engaged in the sale of finished products for consumer end users.

The petrochemical market segment of industrial and small and large companies that invest in oil, gas and chemicals and sell these products to consumers industry.

Peaks in workload demands for the retail market segment are generally predictable (such as time-critical Diwali shopping).

The peaks for the petrochemical market segment are generally based on different factors that are not so easy to predict, however, they generally follow: the economy, the demand for supply chain optimization, investment in oil drilling deep
and climatic conditions Adverse unpredictable (as in warm winter one year, snowstorms in the next).

The differences in the types of transactions (vs. industrial. Retailers) and the option of clouds public, private or hybrid affect the creation of the policy limit.

The transaction types are used to group items of revenue and expenses according to business groups or products.

Good policy of the Internal Boundary

Good resource management is important to balance the resource consumption in the cloud environment. A policy limit ensures that the consumption of resources is dynamically balanced for testing and production applications.

The test application may have different threshold requirements than those for production. Use capacity planning in advance to prepare your system to allocate additional resources to instances when the workload demands of reaching the threshold level.

Although IT professionals are used to thinking in abstract terms, a key aspect to address the creation of the policy limit is to remember that a critical component of workload demands is physical. You depend on the rates of physical component
reliability, even with the wireless bits.

The policy should set out what should be a threshold level as the threshold level of 75% or 85% of the capacity of one or more disks. Should include mechanisms for logging and monitoring of resource consumption.

Besides the capacity, when the level is reached, the number of instances of resources allocated and the response time in the allocation of instances should be in the logs. Moreover, the logs should include:

  • Quality of stateful application
  • Points of resumption
  • Failover mechanism
  • Security in cloud services

Stateful Quality:

Stateful quality refers to the state of the application that responds appropriately to subsequent stages of the application functions in the cloud environment. For example, a state should go to the next state function following this very simplified scenario:

  • The consumer selects an item online retail
  • The retailer places the selected item in the basket of the shopping cart
  • The customer provides credit card information
  • The consumer sends the request
  • The retailer validates the credit card information
  • The retailer provides an order number and estimated delivery time
  • The retailer thanks to consumer demand by
  • The consumer gets an e-mail order confirmation
  • The consumer gets an e-mail that the request is being sent

If the state function step 2 has not gone to stage three, what would cause the problem?

The new application builds broke the logic?

When the system detects the threshold level exceeding the demands of the workload, the threshold level was set at a too high level, so that the remaining resources were insufficient to continue the operation..

Resumption points:

The system must create a recovery point (the varieties planned, and manual installation) at different points of time before a problem occurs in the system.

Snapshots of disks that contain points of recovery should be backed up to disk on the local system and on another disk in a different remote location. The log should indicate the time at which the recovery points were created and what the point of
recovery was used to restore the system.

Engine Fail over:

The system must also be able to initiate failover mechanisms to continue the availability of the operation.

The failover mechanisms must include alternative connecting wired or wireless case, for example, the provider of telecommunication accidentally cut the fiber line or shut down the wireless network connected to the physical installation of the consumer. The log should indicate the type and location of the device used in the failover.


Leave a Reply

Follow by Email