SafePeak - Resolving performance challenges solutions: Hardware upgrades VS In-memory result caching
Hardware upgrade vs. In-memory Caching
Download the whitepaper
Introduction
Medium size companies, large enterprises and municipal or governmental organizations run business applications such as CRM, ERP or document management (CMS), for customer service, resources planing, management and collaboration. Successful management of businesses is in heavy reliance on optimal performance of those corporate information systems.
Be it the length of call waiting queues at a contact center or the number of dropped calls, the wait time of an engineer to get a response while collaborating with a colleague or a CEO waiting for a document before a meeting, it always comes to a quantified performance dollar equivalent. How much could we save or grow our revenues? How much money could the business save if the Sharepoint / CRM was 50% faster? Alternatively, how much more productive could engineers or office clerks be and could it shorten product time to market? Could the collection department be more efficient in achieving its goals or could sales be ramped up quicker with a faster moving CRM?
Businesses report application performance degradation as amount of data grows, usage growth (number of users, number of operating hours, etc.) and specifically during high season peaks or following marketing campaigns. In fact, when multiple threads or processes run on a multi-core / multi-processor servers, they compete for shared resources such as memory and can suffer performance degradation as high as 200%.
It is not surprising that businesses constantly search solutions to improve application performance to grow sales, improve customer satisfaction and achieve operating excellence. The purpose of this paper is to assist businesses choosing a reasonable path to achieving performance improvement goals. It discusses visible and hidden cost issues.
Application and database performance challenges are complicated. Root causes can reside in multiple areas. What are possible problem causes and which solution alternatives should be considered?
Though there is always a trade-off between performance and hardware utilization, it is well known that performance is not linearly related to hardware capacity; that is to say - growing hardware capacity in memory and / or core count and speed aspects doesn’t yield linear performance improvement! So, what are the alternatives? Constant performance tuning by database administrators is one alternative. It keeps DBAs awake at nights and at work during holidays high season periods. Software programing is another alternative. Be it cache or indexes improvement, programmers can spend hours and weeks in optimizing performance. Nevertheless, as we will see in the example of parallelism, this more often than not, provides little help; the mystery of CXPackets.
This whitepaper describes the different aspects required to consider when reviewing alternatives for performance improvement
Download the whitepaper