Skip to main content
In Memoriam

I have been reading a White Paper on cost of ownership and Server memory from Kingston Technology and the results are interesting enough to share.



Most businesses are trying to find ways of saving money on IT spend and drive down cost of ownership. Invariably, an expense that can’t be ignored is that of the Server upgrade. Many companies buy Servers with a dual processor capability but only a relatively small percentage of owners actually choose to add a second CPU or more memory, rather than buy a completely new Server when circumstances demand greater performance. Therefore, when a new Operating System, such as Windows 2003 comes along, there’s a tendency among many business, encouraged by the leading hardware vendors, to replace an existing Server with a shiny new one.

One objective of Kingston’s research, conducted by Mindcraft Labs, was to establish a ‘test-point’, which could aid customers in determining the best and most cost-effective configuration for their Servers and this revealed two very interesting facts.

In tests with Windows 2000, Linux and Solaris, adding more memory can increase Server performance by a much as 50% on a Web Server and that extra memory, rather than a second CPU, commonly represents a more cost-effective solution to the challenge of improving Web Server and DBMS Server performance across operating platforms. As an example, quadrupling the memory in a Windows 2000 Server from 512 MB to 2GB is 40% more effective than simply adding a second processor and the performance improvement could be as high as 1000% on dual processor Solaris DBMS systems with the memory expanded from 512Mb to 4 GB.

The results clearly indicate that adding memory first, as a rule of thumb offers the best solution until an application’s need for memory is completely satisfied. Once this objective has been achieved, then adding further processors is a next step but one should always ensure that systems are fully memory efficient first before adding further processors.

The research also argues against throwing-out older Servers when new applications or Operating Systems appear. In a great many cases, a cheap memory upgrade is enough to maintain the status quo or even improve performance significantly on what it had been in the past. In some respects, it’s rather like buying a new TV because you want to watch BBC Four, when in fact all one needs is the equipment upgrade.

So unless your business really wants to spend its dwindling IT budget on new boxes, it’s worth remembering that memory is the key factor to consider in respect of applications performance. Most of us don’t have enough of it and a little extra can, as the research illustrates, make a significant impact on Web Server or Database Server performance.

The research can be found at http://www.kingston.com/europe/mindcraft/


Comments

Popular posts from this blog

Mainframe to Mobile

Not one of us has a clue what the world will look like in five years’ time, yet we are all preparing for that future – As  computing power has become embedded in everything from our cars and our telephones to our financial markets, technological complexity has eclipsed our ability to comprehend it’s bigger picture impact on the shape of tomorrow.

Our intuition has been formed by a set of experiences and ideas about how things worked during a time when changes were incremental and somewhat predictable. In March 1953. there were only 53 kilobytes of high-speed RAM on the entire planet.

Today, more than 80 per cent of the value of FTSE 500* firms is ‘now dark matter’: the intangible secret recipe of success; the physical stuff companies own and their wages bill accounts for less than 20 per cent: a reversal of the pattern that once prevailed in the 1970s. Very soon, Everything at scale in this world will be managed by algorithms and data and there’s a need for effective platforms for ma…

Civilisational Data Mining

It’s a new expression I haven’t heard before. ‘Civilisational data mining.’

Let me start by putting it in some context. Every character, you or I have typed into the Google search engine or Facebook over the last decade, means something, to someone or perhaps ‘something,’ if it’s an algorithm.


In May 2014, journalists revealed that the United States National Security Agency, the NSA, was recording and archiving every single cell-phone conversation that took place in the Bahamas. In the process they managed to transform a significant proportion of a society’s day to day interactions into unstructured data; valuable information which can of course be analysed, correlated and transformed for whatever purpose the intelligence agency deems fit.

And today, I read that a GOP-hired data company in the United States has ‘leaked’ personal information, preferences and voting intentions on… wait for it… 198 million US citizens.

Within another decade or so, the cost of sequencing the human genome …

The Big Steal

I’m not here to predict the future;” quipped the novelist, Ray Bradbury. “I’m here to prevent it.” And the future looks much like one where giant corporations who hold the most data, the fastest servers, and the greatest processing power will drive all economic growth into the second half of the century.

We live in an unprecedented time. This in the sense that nobody knows what the world will look like in twenty years; one where making confident forecasts in the face of new technologies becomes a real challenge. Before this decade is over, business leaders will face regular and complex decisions about protecting their critical information and systems as more of the existing solutions they have relied upon are exposed as inadequate.

The few real certainties we have available surround the uninterrupted march of Moore’s Law - the notion that the number of transistors in the top-of-the-line processors doubles approximately every two years - and the unpredictability of human nature. Exper…