Proactive and reactive thermal aware optimization techniques to minimize the environmental impact of data centers

Los Centros de Datos se encuentran actualmente en cualquier sector de la economía mundial. Están compuestos por miles de servidores, dando servicio a los usuarios de forma global, las 24 horas del día y los 365 días del año. Durante los últimos años, las aplicaciones del ámbito de la e-Ciencia, como la e-Salud o las Ciudades Inteligentes han experimentado un desarrollo muy significativo. La necesidad de manejar de forma eficiente las necesidades de cómputo de aplicaciones de nueva generación, junto con la creciente demanda de recursos en aplicaciones tradicionales, han facilitado el rápido crecimiento y la proliferación de los Centros de Datos. El principal inconveniente de este aumento de capacidad ha sido el rápido y dramático incremento del consumo energético de estas infraestructuras. En 2010, la factura eléctrica de los Centros de Datos representaba el 1.3% del consumo eléctrico mundial. Sólo en el año 2012, el consumo de potencia de los Centros de Datos creció un 63%, alcanzando los 38GW. En 2013 se estimó un crecimiento de otro 17%, hasta llegar a los 43GW. Además, los Centros de Datos son responsables de más del 2% del total de emisiones de dióxido de carbono a la atmósfera. Esta tesis doctoral se enfrenta al problema energético proponiendo técnicas proactivas y reactivas conscientes de la temperatura y de la energía, que contribuyen a tener Centros de Datos más eficientes. Este trabajo desarrolla modelos de energía y utiliza el conocimiento sobre la demanda energética de la carga de trabajo a ejecutar y de los recursos de computación y refrigeración del Centro de Datos para optimizar el consumo. Además, los Centros de Datos son considerados como un elemento crucial dentro del marco de la aplicación ejecutada, optimizando no sólo el consumo del Centro de Datos sino el consumo energético global de la aplicación. Los principales componentes del consumo en los Centros de Datos son la potencia de computación utilizada por los equipos de IT, y la refrigeración necesaria para mantener los servidores dentro de un rango de temperatura de trabajo que asegure su correcto funcionamiento. Debido a la relación cúbica entre la velocidad de los ventiladores y el consumo de los mismos, las soluciones basadas en el sobre-aprovisionamiento de aire frío al servidor generalmente tienen como resultado ineficiencias energéticas. Por otro lado, temperaturas más elevadas en el procesador llevan a un consumo de fugas mayor, debido a la relación exponencial del consumo de fugas con la temperatura. Además, las características de la carga de trabajo y las políticas de asignación de recursos tienen un impacto importante en los balances entre corriente de fugas y consumo de refrigeración. La primera gran contribución de este trabajo es el desarrollo de modelos de potencia y temperatura que permiten describes estos balances entre corriente de fugas y refrigeración; así como la propuesta de estrategias para minimizar el consumo del servidor por medio de la asignación conjunta de refrigeración y carga desde una perspectiva multivariable. Cuando escalamos a nivel del Centro de Datos, observamos un comportamiento similar en términos del balance entre corrientes de fugas y refrigeración. Conforme aumenta la temperatura de la sala, mejora la eficiencia de la refrigeración. Sin embargo, este incremente de la temperatura de sala provoca un aumento en la temperatura de la CPU y, por tanto, también del consumo de fugas. Además, la dinámica de la sala tiene un comportamiento muy desigual, no equilibrado, debido a la asignación de carga y a la heterogeneidad en el equipamiento de IT. La segunda contribución de esta tesis es la propuesta de técnicas de asigación conscientes de la temperatura y heterogeneidad que permiten optimizar conjuntamente la asignación de tareas y refrigeración a los servidores. Estas estrategias necesitan estar respaldadas por modelos flexibles, que puedan trabajar en tiempo real, para describir el sistema desde un nivel de abstracción alto. Dentro del ámbito de las aplicaciones de nueva generación, las decisiones tomadas en el nivel de aplicación pueden tener un impacto dramático en el consumo energético de niveles de abstracción menores, como por ejemplo, en el Centro de Datos. Es importante considerar las relaciones entre todos los agentes computacionales implicados en el problema, de forma que puedan cooperar para conseguir el objetivo común de reducir el coste energético global del sistema. La tercera contribución de esta tesis es el desarrollo de optimizaciones energéticas para la aplicación global por medio de la evaluación de los costes de ejecutar parte del procesado necesario en otros niveles de abstracción, que van desde los nodos hasta el Centro de Datos, por medio de técnicas de balanceo de carga. Como resumen, el trabajo presentado en esta tesis lleva a cabo contribuciones en el modelado y optimización consciente del consumo por fugas y la refrigeración de servidores; el modelado de los Centros de Datos y el desarrollo de políticas de asignación conscientes de la heterogeneidad; y desarrolla mecanismos para la optimización energética de aplicaciones de nueva generación desde varios niveles de abstracción. ABSTRACT Data centers are easily found in every sector of the worldwide economy. They consist of tens of thousands of servers, serving millions of users globally and 24-7. In the last years, e-Science applications such e-Health or Smart Cities have experienced a significant development. The need to deal efficiently with the computational needs of next-generation applications together with the increasing demand for higher resources in traditional applications has facilitated the rapid proliferation and growing of data centers. A drawback to this capacity growth has been the rapid increase of the energy consumption of these facilities. In 2010, data center electricity represented 1.3% of all the electricity use in the world. In year 2012 alone, global data center power demand grew 63% to 38GW. A further rise of 17% to 43GW was estimated in 2013. Moreover, data centers are responsible for more than 2% of total carbon dioxide emissions. This PhD Thesis addresses the energy challenge by proposing proactive and reactive thermal and energy-aware optimization techniques that contribute to place data centers on a more scalable curve. This work develops energy models and uses the knowledge about the energy demand of the workload to be executed and the computational and cooling resources available at data center to optimize energy consumption. Moreover, data centers are considered as a crucial element within their application framework, optimizing not only the energy consumption of the facility, but the global energy consumption of the application. The main contributors to the energy consumption in a data center are the computing power drawn by IT equipment and the cooling power needed to keep the servers within a certain temperature range that ensures safe operation. Because of the cubic relation of fan power with fan speed, solutions based on over-provisioning cold air into the server usually lead to inefficiencies. On the other hand, higher chip temperatures lead to higher leakage power because of the exponential dependence of leakage on temperature. Moreover, workload characteristics as well as allocation policies also have an important impact on the leakage-cooling tradeoffs. The first key contribution of this work is the development of power and temperature models that accurately describe the leakage-cooling tradeoffs at the server level, and the proposal of strategies to minimize server energy via joint cooling and workload management from a multivariate perspective. When scaling to the data center level, a similar behavior in terms of leakage-temperature tradeoffs can be observed. As room temperature raises, the efficiency of data room cooling units improves. However, as we increase room temperature, CPU temperature raises and so does leakage power. Moreover, the thermal dynamics of a data room exhibit unbalanced patterns due to both the workload allocation and the heterogeneity of computing equipment. The second main contribution is the proposal of thermal- and heterogeneity-aware workload management techniques that jointly optimize the allocation of computation and cooling to servers. These strategies need to be backed up by flexible room level models, able to work on runtime, that describe the system from a high level perspective. Within the framework of next-generation applications, decisions taken at this scope can have a dramatical impact on the energy consumption of lower abstraction levels, i.e. the data center facility. It is important to consider the relationships between all the computational agents involved in the problem, so that they can cooperate to achieve the common goal of reducing energy in the overall system. The third main contribution is the energy optimization of the overall application by evaluating the energy costs of performing part of the processing in any of the different abstraction layers, from the node to the data center, via workload management and off-loading techniques. In summary, the work presented in this PhD Thesis, makes contributions on leakage and cooling aware server modeling and optimization, data center thermal modeling and heterogeneityaware data center resource allocation, and develops mechanisms for the energy optimization for next-generation applications from a multi-layer perspective.

Developmental exposure to vasopressin increases aggression in adult prairie voles

Although the biological roots of aggression have been the source of intense debate, the precise physiological mechanisms responsible for aggression remain poorly understood. In most species, aggression is more common in males than females; thus, gonadal hormones have been a focal point for research in this field. Although gonadal hormones have been shown to influence the expression of aggression, in many cases aggression can continue after castration, indicating that testicular steroids are not completely essential for the expression of aggression. Recently, the mammalian neuropeptide arginine vasopressin (AVP) has been implicated in aggression. AVP plays a particularly important role in social behavior in monogamous mammals, such as prairie voles (Microtus ochrogaster). In turn, the effects of social experiences may be mediated by neuropeptides, including AVP. For example, sexually naïve prairie voles are rarely aggressive. However, 24 h after the onset of mating, males of this species become significantly aggressive toward strangers. Likewise, in adult male prairie voles, central (intracerebroventricular) injections of AVP can significantly increase intermale aggression, suggesting a role for AVP in the expression of postcopulatory aggression in adult male prairie voles. In this paper, we demonstrate that early postnatal exposure to AVP can have long-lasting effects on the tendency to show aggression, producing levels of aggression in sexually naïve, adult male prairie voles that are comparable to those levels observed after mating. Females showed less aggression and were less responsive to exogenous AVP, but the capacity of an AVP V1a receptor antagonist to block female aggression also implicates AVP in the development of female aggression.

Proactive and reactive inhibition during overt and covert actions. An electrical neuroimaging study.

Response inhibition is the ability to suppress inadequate but automatically activated, prepotent or ongoing response tendencies. In the framework of motor inhibition, two distinct operating strategies have been described: “proactive” and “reactive” control modes. In the proactive modality, inhibition is recruited in advance by predictive signals, and actively maintained before its enactment. Conversely, in the reactive control mode, inhibition is phasically enacted after the detection of the inhibitory signal. To date, ample evidence points to a core cerebral network for reactive inhibition comprising the right inferior frontal gyrus (rIFG), the presupplementary motor area (pre-SMA) and the basal ganglia (BG). Moreover, fMRI studies showed that cerebral activations during proactive and reactive inhibition largely overlap. These findings suggest that at least part of the neural network for reactive inhibition is recruited in advance, priming cortical regions in preparation for the upcoming inhibition. So far, proactive and reactive inhibitory mechanisms have been investigated during tasks in which the requested response to be stopped or withheld was an “overt” action execution (AE) (i.e., a movement effectively performed). Nevertheless, inhibitory mechanisms are also relevant for motor control during “covert actions” (i.e., potential motor acts not overtly performed), such as motor imagery (MI). MI is the conscious, voluntary mental rehearsal of action representations without any overt movement. Previous studies revealed a substantial overlap of activated motor-related brain networks in premotor, parietal and subcortical regions during overtly executed and imagined movements. Notwithstanding this evidence for a shared set of cerebral regions involved in encoding actions, whether or not those actions are effectively executed, the neural bases of motor inhibition during MI, preventing covert action from being overtly performed, in spite of the activation of the motor system, remain to be fully clarified. Taking into account this background, we performed a high density EEG study evaluating cerebral mechanisms and their related sources elicited during two types of cued Go/NoGo task, requiring the execution or withholding of an overt (Go) or a covert (MI) action, respectively. The EEG analyses were performed in two steps, with different aims: 1) Analysis of the “response phase” of the cued overt and covert Go/NoGo tasks, for the evaluation of reactive inhibitory control of overt and covert actions. 2) Analysis of the “preparatory phase” of the cued overt and covert Go/NoGo EEG datasets, focusing on cerebral activities time-locked to the preparatory signals, for the evaluation of proactive inhibitory mechanisms and their related neural sources. For these purposes, a spatiotemporal analysis of the scalp electric fields was applied on the EEG data recorded during the overt and covert Go/NoGo tasks. The spatiotemporal approach provide an objective definition of time windows for source analysis, relying on the statistical proof that the electric fields are different and thus generated by different neural sources. The analysis of the “response phase” revealed that key nodes of the inhibitory circuit, underpinning inhibition of the overt movement during the NoGo response, were also activated during the MI enactment. In both cases, inhibition relied on the activation of pre-SMA and rIFG, but with different temporal patterns of activation in accord with the intended “covert” or “overt” modality of motor performance. During the NoGo condition, the pre-SMA and rIFG were sequentially activated, pointing to an early decisional role of pre-SMA and to a later role of rIFG in the enactment of inhibitory control of the overt action. Conversely, a concomitant activation of pre-SMA and rIFG emerged during the imagined motor response. This latter finding suggested that an inhibitory mechanism (likely underpinned by the rIFG), could be prewired into a prepared “covert modality” of motor response, as an intrinsic component of the MI enactment. This mechanism would allow the rehearsal of the imagined motor representations, without any overt movement. The analyses of the “preparatory phase”, confirmed in both overt and covert Go/NoGo tasks the priming of cerebral regions pertaining to putative inhibitory network, reactively triggered in the following response phase. Nonetheless, differences in the preparatory strategies between the two tasks emerged, depending on the intended “overt” or “covert” modality of the possible incoming motor response. During the preparation of the overt Go/NoGo task, the cue primed the possible overt response programs in motor and premotor cortex. At the same time, through preactivation of a pre-SMA-related decisional mechanism, it triggered a parallel preparation for the successful response selection and/or inhibition during the subsequent response phase. Conversely, the preparatory strategy for the covert Go/NoGo task was centred on the goal-oriented priming of an inhibitory mechanism related to the rIFG that, being tuned to the instructed covert modality of the motor performance and instantiated during the subsequent MI enactment, allowed the imagined response to remain a potential motor act. Taken together, the results of the present study demonstrate a substantial overlap of cerebral networks activated during proactive recruitment and subsequent reactive enactment of motor inhibition in both overt and covert actions. At the same time, our data show that preparatory cues predisposed ab initio a different organization of the cerebral areas (in particular of the pre-SMA and rIFG) involved with sensorimotor transformations and motor inhibitory control for executed and imagined actions. During the preparatory phases of our cued overt and covert Go/NoGo tasks, the different adopted strategies were tuned to the “how” of the motor performance, reflecting the intended overt and covert modality of the possible incoming action.

The Proactive International Dimension and the Breakdown of Authoritarian Regimes

There is international consensus among scholars that democratic transitions are multicausal processes in which both internal and international variables are involved (Pridham 1991, 1995; Whitehead 1996; Schmitter 1996; Linz and Stepan 1996; Carothers 1999; Morlino and Magen 2008; Grilli di Cortona 2009). This chapter is limited, on the one hand, to the dependent variable consisting solely of the crisis/breakdown/transformation of non-democratic regimes in the Third Wave of democratization, and, on the other hand, to an independent variable identified solely with the international dimension of democratic transition. This factor, which can be termed the Proactive International Dimension (PID), specifically concerns that combination of actions or processes, produced by one or more international actors, that, intentionally or not, cause or contribute to the crisis/breakdown/transformation of a non-democratic regime.

Driver Aggression as a Function of Status Concurrence: An Analysis of Horn-Honking Responses

Various field experiments were conducted to examine the influence of social status on aggression in road traffic. Horn-honking response times of subjects blocked by an experimental car at traffic lights were considered to be an indicator of the degree of aggression. During an initial experiment, the status of the frustrator was varied and an inverse relation was observed between status and aggression towards the frustrator. On the other hand, in a more recent experiment higher status aggressors were found to behave more aggressively. In our study we combined the two designs, i.e., we varied the status of the frustrator and at the same time measured the status of the aggressor. Neither results of the former experiments could be replicated, but we observed a reduction in aggression when frustrator and aggressor were of similar social status.

Masking odour when regrouping rabbit does: Effect on aggression, stress and lesions.

Regrouping female rabbits (Oryctolagus cuniculus) in group housing systems is a common management practice in Swiss rabbit breeding which may, however, induce agonistic interactions resulting in social stress and severe lesions. On farms using artificial insemination, does are usually kept singly for 12 days after parturition to avoid pseudopregnancy and fighting for nests. The integration of new group members usually occurs after this isolation phase. This study was conducted with 128 gravid does of the Hycole hybrid, housed in pens covering a floor area of 5.7 m2 that was bedded with straw and furnished with elevated areas, hiding places and eight compartments with nest boxes. In the experiment, the fur of 16 groups of 8 does each was sprayed with either alcohol or vinegar to mask the pre-existing group odours, or with water (control groups) shortly before regrouping. Lesion scores, stress parameters (body temperature and blood glucose level) and behaviour were assessed before and after the isolation phase. Effects of treatment and time on all collected parameters were analysed using mixed models. On the second day after regrouping 43% of the does showed new lesions. In the first five days after regrouping, new lesions occurred in 60% of the does; 32% had severe lesions. After regrouping, more agonistic interactions were observed and body temperature and blood glucose levels were higher than before regrouping (P<0.001 each). Body temperature increased less in groups treated with vinegar compared to the other two treatments on the first day after regrouping (P=0.017). In all other parameters no influence of the treatment with alcohol or vinegar was found. These findings suggest that masking the group odours with alcohol or vinegar had little effect on lesions, stress and agonistic interactions. Therefore, alternative management procedures need to be developed to reduce lesions and stress caused by aggressive behaviour.

Nazi conspiracy and aggression.

Mode of access: Internet.

Nazi conspiracy and aggression, opinion and judgment. Office of United States Chief of Counsel for Prosecution of axis criminality.

Seals of Dept. of State, War Dept., and International Military Tribunal, on t.p.

A proposal for a total peace offensive : to stop Communist aggression by taking the initiative in the world contest for men's minds, hearts and loyalties /

Mode of access: Internet.

Of cats and rats : studies of the neural basis of aggression /

Mode of access: Internet.

Nazi conspiracy and aggression.

"Biographical index of principal persons referred to in the interrogations": Supplement B, p. 1679-1713.

Libya under Qadhafi : a pattern of aggression

Caption title.

State of Connecticut by His Excellency John H. Trumbull governor a proclamation : It is no small part of the honorable record of the United States that every war in which we have been engaged has been a struggle, not of greed and aggression, but of defence or deliverance ... pursuant to the statute which properly directs a day to be set aside in memory of the tragic event which precipitated the Spanish War, I designate Monday, February fifteenth next, as Maine Memorial Day ... given under my hand and seal of the state at the Capitol, in Hartford, this twenty-sixth day of January ... one thousand nine hundred and twenty-six ...

Arms of the state of Connecticut printed at head and state seal printed at lower left.