Server consolidation is currently widely employed in order to improve the energy efficiency of data centers. While being a promising technique, server consolidation may lead to resource interference between applications and thus, reduced performance of applications. Current approaches to account for possible resource interference are not well suited to respect the variation in the workloads for the applications. As a consequence, these approaches cannot prevent resource interference if workload for applications vary. It is assumed that having models for the resource utilization and power consumption of applications as functions of the workload to the applications can improve decision making and help to prevent resource interference in scenarios with varying workload. This thesis aims to develop such models for selected applications. To produce varying workload that resembles statistical properties of real-world workload a workload generator is developed in a first step. Usually, the measurement data for such models origins from different sensors and equipment, all producing data at different frequencies. In order to account for these different frequencies, in a second step this thesis particularly investigates the feasibility to employ quantile functions as model inputs. Complementary, since conventional goodness-of-fit tests are not appropriate for this approach, an alternative to assess the estimation error is presented.:1 Introduction 2 Thesis Overview 2.1 Testbed 2.2 Contributions and Thesis Structure 2.3 Scope, Assumptions, and Limitations 3 Generation of Realistic Workload 3.1 Statistical Properties of Internet Traffic 3.2 Statistical Properties of Video Server Traffic 3.3 Implementation of Workload Generation 3.4 Summary 4 Models for Resource Utilization and for Power Consumption 4.1 Introduction 4.2 Prior Work 4.3 Test Cases 4.4 Applying Regression To Samples Of Different Length 4.5 Models for Resource Utilization as Function of Request Size 4.6 Models for Power Consumption as Function of Resource Utilization 4.7 Summary 5 Conclusion & Future Work 5.1 Summary 5.2 Future Work Appendices / Serverkonsolidierung wird derzeit weithin zur Verbesserung der Energieeffizienz von Rechenzentren eingesetzt. Während diese Technik vielversprechende Ergebnisse zeitigt, kann sie zu Ressourceninterferenz und somit zu verringerter Performanz von Anwendungen führen. Derzeitige Ansätze, um dieses Problem zu adressieren, sind nicht gut für Szenarien geeignet, in denen die Workload für die Anwendungen variiert. Als Konsequenz daraus folgt, dass diese Ansätze Ressourceninterferenz in solchen Szenarien nicht verhindern können. Es wird angenommen, dass Modelle für Anwendungen, die deren Ressourenauslastung und die Leistungsaufnahme als Funktion der Workload beschreiben, die Entscheidungsfindung bei der Konsolidierung verbessern und Ressourceninterferenz verhindern können. Diese Arbeit zielt darauf ab, solche Modelle für ausgewählte Anwendungen zu entwickeln. Um variierende Workload zu erzeugen, welche den statistischen Eigenschaften realer Workload folgt, wird zunächst ein Workload-Generator entwickelt. Gewöhnlicherweise stammen Messdaten für die Modelle aus verschienenen Sensoren und Messgeräten, welche jeweils mit unterschiedlichen Frequenzen Daten erzeugen. Um diesen verschiedenen Frequenzen Rechnung zu tragen, untersucht diese Arbeit insbesondere die Möglichkeit, Quantilfunktionen als Eingabeparameter für die Modelle zu verwenden. Da konventionelle Anpassungsgütetests bei diesem Ansatz ungeeignet sind, wird ergänzend eine Alternative vorgestellt, um den durch die Modellierung entstehenden Schätzfehler zu bemessen.:1 Introduction 2 Thesis Overview 2.1 Testbed 2.2 Contributions and Thesis Structure 2.3 Scope, Assumptions, and Limitations 3 Generation of Realistic Workload 3.1 Statistical Properties of Internet Traffic 3.2 Statistical Properties of Video Server Traffic 3.3 Implementation of Workload Generation 3.4 Summary 4 Models for Resource Utilization and for Power Consumption 4.1 Introduction 4.2 Prior Work 4.3 Test Cases 4.4 Applying Regression To Samples Of Different Length 4.5 Models for Resource Utilization as Function of Request Size 4.6 Models for Power Consumption as Function of Resource Utilization 4.7 Summary 5 Conclusion & Future Work 5.1 Summary 5.2 Future Work Appendices

Function Shift is the transference of functions, which involves responsibilities, assets and human resources (including their employment packages), from one department to the next. The Function Shift to which I refer in this study entails the shifting of functions from the former Adult Education and Training provincial directorates to the Department of Higher Education and Training (DHET). This process started in 2009 in terms of proclamation 48 of 2009. The purpose of my active research is to investigate in depth the experienced positive and negative consequences of Function Shift with the intention of exploring problematic features and challenges of Community Education and possibilities for addressing them. By ‘experienced’ consequences, I mean consequences that are not just imagined but were expressed by participants. I employed an 'active' qualitative research approach whereby I, as a researcher, am actively involved in the research process in trying to ensure that the research is bearing results for me as well as for the participants. I based the selection of Community Education and Training Colleges on the characteristics of the regions in which the colleges belonged. I clustered regions that portrayed similar characteristics and came out with 3 clusters. I selected one region and its respective college from each of the 3 clusters. From each of the selected regions and their corresponding colleges, I sampled a Regional manager, Curriculum Implementer or regional official, Principal, 1 Centre manager, 1 lecturer and 1 student. I collected data through first and second interview sessions, focus group discussion in 1 college and through evaluative discussion with 2 head office officials. To carry out data analysis, I used the principles of Atlas TI that encourages the coding, categorisation and thematising data collected from participants simultaneously with data from the reviewed literature. It emerged that all participants agreed that a multilevel change management system is suitable for Function Shift as opposed to the traditional rational/linear model and that Function Shift is a potential solution to Adult Education and Training challenges. The prevailing challenge was insufficient consultation, which resulted in some transitional challenges that could have been identified and mitigated against. My concluding recommendation is that the oral or print input made by members on the ground including the assessment of the real and practical situation in Community Learning Centres must drive the development of policies that are still cascaded by the DHET. Consultation must be characterised by dialogue, not announcements of deadlines. / Ku susumetiwa ka mintirho swi vula ku susiwa ka vutihlamuleri endzawuleni yinwana byi yisiwa endzawuleni yin’wani. Vutihlamuleri lebyi byi katsa tinhundu, timali, vatirhi ni miholo ya vona ni hikwaswo leswi fambelanaka ni xiyenge xexo. Ndzavisiso lowu wu vulavula hi ku susiwa ka vutihlamuleri bya dyondzo ni vudzaberi/vuthwaseli bya vatswatsi (Adult Education and Training) e mindzawuleni ya dyondzo ya le hansi ya swifunda (Provincial Department of Basic Education) ku yisiwa e ndzawuleni ya le henhla ya dyondzo ni vudzaberi (Department of Higher Education and Training). Nghingiriko lowu wa ku cinciwa ka vutihlamuleri wu sungurile hi lembe ra 2009. Makungu ya ndzavisiso lowu wa mahika I ku lavisisa hi vuxokoxoko vumbhoni bya switandzaku (mbuyelo lowunene ni lowu wu nga tsakisiki) leswi vangiwanga hi ku cinciwa ka vutihlamuleri, hi xikongomelo xo paluxa swirhalanganyi swa dyondzo ya vaaki (Community Education) ni ku ololoxa swirhalanganyi leswi. Loko ni ku vumbhoni bya switandzaku, ndzi vula switandzaku leswi swi nga kumbeteriwiki, kambe leswi vahlamuri (participants/respondents) va nyikaka vumbhoni bya leswi va nga swi vona ni ku switwa. Ndzi endlile vulavisisi bya mahika (active research), laha mina tani hi mulavisisi ndzi tlangeke xiyenge xa ku endla leswaku vulavisisi lebyi byi va ni mbuyelo lowu nga ta pfuna mina xikan’we na muhlamuri. Ndzi hlawurile tilholichi ta dyondzo ni vudzaberi ta vaaki ku ya hi tindhawu /tirhijini laha tikholichi leti ti kumekaka kona. Ndzi longoloxile tirhijini hinkwato, ndzi ti katsakanya hi timpawu ta tona, ivi ndzi huma na mintlawa minarhu. Ndzi hlawule kholeji yin’we eka ntlawa wun’wani ni wun’wani ni tirhijini ta tona. Eka rhijini yin'wana na yin'wana ndzi hlawurile no tihlanganisa na vanhu lava landzelaka: mufambisi wa rhijini, mukamberi/museketeri wa dyondzo a rhijinini, nhloko ya kholeji, mufambisi wa sentara, mudzaberi na xichudeni. Eka Kholeji yo sungula ni ya vumbirhi, ndzi hlengeletile mahungu hi ku burisana ni vahlamuri hi wun’we ha wun’we. Eka Kholeji ya vunharhu, ndzi hlengelete mahungu hi mbhurisano wa hlengeletano ya murhangeri wa senthara, vadzaberi va nharhu ni machudeni mambirhi. Ku kuma voxokoxoko ni nhlavutelo wa mahungu lawa ndzi wa hlengeleteke, ndzi tirhisile maendlelo ya "Atlas Tl" yaku hlohlotela ku kuma vuxokoxoko hi ku tirhisa tekinoloji, ku longoloxa ku ya hi swiyimo ni ku endla vulavisisi eka tibuku tin'wana. Vahlamuri va pfumelelanile leswaku mafambiselo ya ku cinca loku khumbhaka swiyenge swo hambana-hambana (Multilevel change management) hi nkarhi wun’we hi wona lama fanelaka ku susumetiwa ka vutihlamuleri. Nakambe vahlamuri va pfumelelanile leswaku ku susmetiwa ka vutihlamuleri swi nga tisa xintshuxo eka ku tikeriwa loku a ku ri kona e ka dyondzo ni vudzaberi bya vatswatsi. Ndzi heta hi ku vula leswaku swibumabumelo leswi tsariweke ni ku vuriwa hi milomo ya vaaki, ni ku xopaxopela xiyimo lexi xi nga etisenthareni ta dyondzo ya vaaki, hi swona leswi fanelaka ku va makombandlela ya ku tumbuluxiwa ni ku hangalasiwa ka milawu (policies) leyi ya ha endliwaka hi ndzawulo ya le henhla ya dyondzo ni vudzaberi. Njhenhjekisano wa miehleketo exikarhi ka varhangeri ni vaaki hi yona ndlela ya kahle yaku tihlanganisa (consultation) na vanhu. / Phetišetšo ya mošomo ke go fetišetša mešomo, yeo e amago maikarabelo, dithoto le methopo ya batho (go akaretšwa ditshwanelo tša bona tša mošomo), go tloga go kgoro ye nngwe go ya go ye nngwe. Phetišetšo ya mošomo yeo ke bolelago ka yona ka mo dinyakišišong e ama go fetišetša mešomo ya Thuto ya Batho ba Bagolo le Tlhahlo go tloga go diofisi tša bolaodibogolo bja diprofense tša Thuto ya Batho ba Bagolo le Tlhahlo tša pele go ya go go Thuto ya Godimo le Tlhahlo (DHET). Tshepetšo ye e thomile ka 2009 go ya ka pego ya 48 ya 2009. Nepo ya dinyakišišo tša ka tša go rarolla bothata ke go nyakišiša go tsenelela dipoelo tše dibotse le tše dimpe tša maitemogelo tša Phetišetšo ya Mošomo ka nepo ya go nyakišiša dibopego tša mathata le ditlhohlo tša Thuto ya Setšhaba le dikgonagalo tša go šogana le tšona. Ka ‘dipoelo tša maitemogelo’ ke bolela ka dipoelo tšeo di sa akanywego fela eupša di tšweletšwa ke bakgathatema. Ke šomišitše mokgwa wa dinyakišišo wa boleng wa “go rarolla bothata’ moo nna, bjalo ka monyakišiši, ke amana ka dinyakišišong ka mafolofolo go kgonthiša gore dinyakišišo di na le dipoelo tše dibotse go nna le go bakgathatema. Ke theile kgetho ya ka ya Thuto ya Setšhaba le Dikholetšhe tša Tlhahlo go dibopego tša dilete tšeo dikholetšhe tše di lego gona. Ke hlopile dilete tšeo di bontšhago dibopego tša go swana gomme ka tšweletša dihlopha tše tharo. Go tšwa go dilete tše dingwe le tše dingwe tšeo di kgethilwego le dikholetšhe tšeo di amanago le tšona, ke dirile sešupo ka molaodi wa Selete, Mophethagatši wa Lenaneothuto goba mohlankedi wa selete, Hlogo ya Sekolo, molaodi wa Senthara yo motee, mofahloši yo motee le moithuti yo motee. Ke kgobokeditše data ka dikopano tša mathomo le tša bobedi tša dipoledišano, dipoledišano tša sehlopha sa nepišo kholetšheng ye tee ka dipoledišano tša tekolo le bahlankedi ba babedi ba kantorokgolo. Go dira tshekatsheko ya data, ke šomišitše methopo ya Atlas TI ye e hlohleletšago go swaya, go hlopha le go kgetha data ye e kgobokeditšwego go tšwa go bakgathatema ka nako ye tee le data go tšwa go dingwalwa tšeo di sekasekilwego. Go tšweletše gore bakgathatema ka moka ba dumetše gore mokgwa wa taolo ya phetogo ya magato a mantši o loketše Phetišetšo ya Mošomo kgahlanong le mmotlolo wa tlwaelo/thwii wa mathomong le gore Phetišetšo ya Mošomo ke tharollo ye e kgonagalago ya ditlhohlo tša Thuto ya Batho ba Bagolo le Tlhahlo. Tlhohlo ye e tšwelelago e be e le therišano yeo e sa lekanago, yeo e feleditšego ka ditlhohlo tša phetišetšo tšeo di bego di utollotšwe gomme tša fedišwa. Tigelo ya ka ya go phetha ke dikgopolo tša molomo le tšeo di gatišitšwego tšeo di filwego ke maloko a mo fase go akaretšwa kelo ya maemo a nnete le a tiro mo Disenthareng tša Go ithuta tša Setšhaba di swanetše go eta pele tšweletšopele ya dipholisi tšeo di sa fetišwago ke DHET. Ditherišano di swanetše go bopša ke poledišano, e sego ditsebišo tša matšatši a mafelelo. / ABET and Youth Development / D. Ed. (Socio Education)

The aim of this study was to investigate the management of the function shift of Technical and Vocational Education and Training Colleges at the Department of Higher Education and Training. When the function shift of TVET Colleges took place, it provided an opportunity to research a phenomenon and obtain data that were not previously researched, as this was the first ever function shift or migration of colleges from the Department of Basic Education to the Department of Higher Education and Training. As such, this research not only explores a new phenomenon but also contributes to the body of knowledge regarding TVET in South Africa. This research was conducted at a TVET College in Newcastle, Kwa-Zulu Natal, South Africa. A qualitative, phenomenological study was conducted. The instruments for collecting data were individual and focus group interviews. Field notes were compiled during the various interviews, as deemed necessary by the researcher. All interviews were recorded digitally. The said digital recordings were transcribed and emanating themes were identified. Managers were selected as the interviewees. As this research was based on the management of the function shift, they were the most appropriate choice. A sample of staff was also interviewed, as two separate focus groups, to balance the scope and extent of the data, thus attempting to not only view the managerial paradigm as the sole reality. Findings made from the data were that the function shift was preceded by a previous migration of staff. That migration, however, saw colleges remaining in the same educational stream or level namely that of basic education. The staff at colleges, or technical colleges as they were called at the time, were previously employed by the State. This changed when the State created governing bodies for colleges or college councils, as they are known. Technical Colleges were then renamed Further Education and Training Colleges. These colleges were reporting to the Department of Education which saw a name change to the Department of Basic Education, as explained above. Employees were given the opportunity to migrate from State employment to council employment, which most staff members did. These college councils were later found to be dysfunctional, and the State arranged for colleges to be returned to the auspices of the State, but this time round they were usurped in the newly formed Department of Higher Education and Training. From a management perspective, the function shift was successful with minor challenges. Thus the staff felt that they did not receive adequate communication, thereby adversely affecting their financial positions. The researcher recommends further research among TVET Colleges regarding the management of this function shift. It is also recommended that all current channels of communication with staff be revisited, and improvement therein pursued, where possible. A final recommendation is that managers involved in a function shift be subject to a refresher course in change management before a function shift is implemented. / Educational Leadership and Management / M. Ed. (Education Management)