8 EFFICIENCY IN COMPUTER SYSTEMS
‘Hiding the hardware’: Using virtual machines to improve system efficiency
Good system design has always tried to balance performance and cost. For designers of generalized commercial systems, this has meant a careful analysis of the specifications of component parts, to ensure that they can provide the necessary functionality and reliability at the least possible cost. Where systems are a key element of commercial organizations, the level of system efficiency can mean the difference between success and failure for a company.
Here is a list of questions which have to be considered
by systems designers: ● What can be done to minimize system hardware costs?
● What is the best balance between the level of specification of hardware components in a system, and their cost?
● What are the implications of these changes on the overall operation of the system?
● How easily will the system architecture ‘scale up’: that is, increase proportionally, if it needs to be expanded?
● How can systems be designed to minimize their effects on the environment?
Some answers to the above questions have been
provided by virtualization. A key concept of virtualization is clustering, which means locating the hardware elements of the system together in a data centre or server farm. This makes it easier to provide an optimal environment for the hardware to work in. System maintenance is also simplified by clustering, as it makes components easily accessible for repair or replacement. Another concept is masking, which involves making the physical components of the system appear as one virtual device to system administrators. This makes it easier for them to manage the system. Virtualization can be used for servers, storage and networking. A key benefit of virtualization is that physical components can be added or removed without shutting down the system (this is known as ‘hot swapping’). This provides great flexibility for system designers as it means that processing, networking or storage capacity can be scaled up or down very quickly in response to changes in the business environment.
In addition to flexibility in system size, virtualization allows flexibility in terms of the specification of system components which can be used, although there are limitations. The system developer’s choice of components is limited by the level of reliability required by the system. For example, ‘out-of-the-box’ servers, which are those purchased from manufacturers without modification, are typically both expensive and high specification. They are ideal for e-commerce companies such as Amazon which require very high levels of reliability and data integrity (ensuring that data is not
corrupted). For others, such as Google, reliability is less important than cost because the amount of revenue which they obtain from each search is so small. For this reason, they have used enormous numbers of low-cost, low-specification PCs. These are less reliable, but using virtualization means that the failure of a component, such as a motherboard, will not disrupt the service overall. Any loss of data will only affect transactions taking place at the time. For search engine systems such as Google, this may simply mean users not getting the optimum search results. For other types of systems, the effects may be more damaging.
Large amounts of power are required to cool data
centres, which has led to a widespread recognition of the impact of virtualization on global warming. The size of this impact can be seen in a 2006 US EPA (Environmental Protection Agency) report. This stated that data centres accounted for 1.5 % of all US electricity consumption, and that the technology they used had to be improved if their rate of growth was to be sustained. For example, most data centres use chillers, which are elaborate water- based cooling systems, to maintain an appropriate temperature for the hardware. However, some companies, such as Google, have begun an initiative to avoid the use of chillers and to reduce power consumption by relying on innovative data centre design. One approach is to use only DC current in the centres, avoiding conversion losses from standard AC current. By using higher specification voltage regulator circuitry – circuits on the computer’s motherboard providing the voltages required by the microchips – more power savings can be made. In fact, all server components are designed to have a property called energy proportionality, which ensures that they operate efficiently whether they are idle (not doing any active work), operating at lower usage levels, or at full capacity. By giving each server its own battery, Google avoids using uninterruptible power supplies (UPS), a term used for giant batteries which ensure that servers keep running if there is a power outage. These consume large amounts of energy. While Google has long been recognized as a leader in data centre energy efficiency, Yahoo is catching up. By 2014, Yahoo plans to have data centres which will be more energy efficient than those currently used by Google, according to a Yahoo corporate blog on 30th June 2009.
65
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73 |
Page 74 |
Page 75 |
Page 76 |
Page 77 |
Page 78 |
Page 79 |
Page 80 |
Page 81 |
Page 82 |
Page 83 |
Page 84 |
Page 85 |
Page 86 |
Page 87 |
Page 88 |
Page 89 |
Page 90 |
Page 91 |
Page 92 |
Page 93 |
Page 94 |
Page 95 |
Page 96 |
Page 97 |
Page 98 |
Page 99 |
Page 100 |
Page 101 |
Page 102 |
Page 103 |
Page 104 |
Page 105 |
Page 106 |
Page 107 |
Page 108 |
Page 109 |
Page 110 |
Page 111 |
Page 112 |
Page 113 |
Page 114 |
Page 115 |
Page 116 |
Page 117 |
Page 118 |
Page 119 |
Page 120 |
Page 121 |
Page 122 |
Page 123 |
Page 124 |
Page 125 |
Page 126 |
Page 127 |
Page 128 |
Page 129 |
Page 130 |
Page 131 |
Page 132 |
Page 133 |
Page 134 |
Page 135 |
Page 136 |
Page 137 |
Page 138 |
Page 139 |
Page 140
