TWM583564U - Cloud resource management system - Google Patents
Cloud resource management system Download PDFInfo
- Publication number
- TWM583564U TWM583564U TW108206991U TW108206991U TWM583564U TW M583564 U TWM583564 U TW M583564U TW 108206991 U TW108206991 U TW 108206991U TW 108206991 U TW108206991 U TW 108206991U TW M583564 U TWM583564 U TW M583564U
- Authority
- TW
- Taiwan
- Prior art keywords
- cloud
- resource
- cloud resource
- processor
- management system
- Prior art date
Links
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
Description
本揭示文件是有關於一種管理系統,且特別是有關於一種雲端資源管理系統。 The present disclosure relates to a management system, and more particularly to a cloud resource management system.
一般來說,企業或服務提供者若需要透過網際網路來提供企業服務或營業時,往往需要花費成本建置硬體資源,以於硬體資源上提供服務。隨著雲端運算技術的發展,服務提供者可運用雲端資源來取代過往的硬體建置,使得服務提供者不再需要花費太高的硬體成本。 Generally speaking, when an enterprise or service provider needs to provide enterprise services or business through the Internet, it often needs to spend the cost of building hardware resources to provide services on hardware resources. With the development of cloud computing technology, service providers can use cloud resources to replace past hardware construction, so that service providers no longer need to spend too high hardware costs.
另一方面,傳統的作法上,服務提供者不僅需要花費硬體成本,在硬體資源的維護方面亦在成本上佔有一定程度的成本花費。因此,轉而在雲端資源上提供服務已成為許多服務提供者的選擇之一。 On the other hand, in the traditional practice, the service provider not only needs to spend hardware costs, but also occupies a certain amount of costs in terms of maintenance of hardware resources. Therefore, switching to providing services on cloud resources has become one of the choices of many service providers.
受惠於雲端服務的多元發展,使得雲端服務相當多樣。然而,對於使用者而言,往往難以於多個雲環境之間選擇欲使用的雲端服務。有鑑於此,如何在多元的雲端資源當中產生雲端服務策略以最大化效益是 亟需解決的技術問題。 Benefiting from the diversified development of cloud services, cloud services are quite diverse. However, for users, it is often difficult to choose a cloud service to use between multiple cloud environments. In view of this, how to generate cloud service strategies among multiple cloud resources to maximize benefits is Technical problems that need to be solved urgently.
根據本揭示文件之一實施例,揭示一種雲端資源管理系統,適配於介接複數個雲環境,各該雲環境包含複數個雲端資源。雲端資源管理系統包含網路介面及處理器。網路介面用以提供雲端資源管理系統與該些複數個雲端資源介接之一介面。處理器耦接網路介面,其中處理器用以:辨識一指令,其中指令定義一雲端資源需求;經由網路介面自該些雲環境獲得關聯於指令之複數個雲端資源需求項目,其中各該雲端資源需求項目包含一參數值,其中參數值包含一資源種類資訊、一資源規格資訊、一資源數量資訊、一資源時間資訊、一資源價格資訊、一收費方式資訊、一資源狀態資訊中至少一者;根據指令針對該些雲端資源需求項目之該些參數值計算出複數個資源組合,其中各資源組合包含該些雲端資源之至少一者;自該些資源組合選定一第一資源組合,第一資源組合包含一第一雲端資源及一第二雲端資源,第一資源組合滿足指令之雲端資源配置需求,其中,第一雲端資源屬於該些雲環境中之一第一雲環境,第二雲端資源屬於該些雲環境中之一第二雲環境;經由網路介面向第一雲環境送出一第一請求,以於第一雲環境中配置第一雲 端資源;以及經由網路介面向第二雲環境送出一第二請求,以於第二雲環境中配置第二雲端資源。 According to an embodiment of the present disclosure, a cloud resource management system is disclosed, adapted to interface a plurality of cloud environments, each of which includes a plurality of cloud resources. The cloud resource management system includes a network interface and a processor. The network interface is used to provide an interface between the cloud resource management system and the plurality of cloud resource interfaces. The processor is coupled to a network interface, where the processor is used to: identify an instruction, wherein the instruction defines a cloud resource requirement; and obtain a plurality of cloud resource requirement items associated with the instruction from the cloud environments via the network interface, each of the cloud The resource requirement item includes a parameter value, where the parameter value includes at least one of a resource type information, a resource specification information, a resource quantity information, a resource time information, a resource price information, a charging method information, and a resource status information. Calculate a plurality of resource combinations according to the instruction for the parameter values of the cloud resource demand items, where each resource combination includes at least one of the cloud resources; a first resource combination is selected from the resource combinations, the first The resource combination includes a first cloud resource and a second cloud resource. The first resource combination satisfies the commanded cloud resource allocation requirements. Among them, the first cloud resource belongs to one of the cloud environments. The first cloud environment and the second cloud resource Belongs to one of these cloud environments, the second cloud environment; sends to the first cloud environment through the network interface A first request to a first configuration of the first cloud in cloud environment End resource; and sending a second request to the second cloud environment via a network interface to configure the second cloud resource in the second cloud environment.
為使本創作的上述技術特徵和優點能更明顯易懂,下文特舉實施例,並配合所附圖式作詳細說明如下。 In order to make the above technical features and advantages of this creation more comprehensible, embodiments are described below in detail with the accompanying drawings as follows.
為讓本揭示內容之上述和其他目的、特徵、優點與實施例能更明顯易懂,所附符號之說明如下: In order to make the above and other objects, features, advantages, and embodiments of the present disclosure more comprehensible, the description of the attached symbols is as follows:
100‧‧‧雲端資源管理系統 100‧‧‧ Cloud Resource Management System
110‧‧‧網路介面 110‧‧‧Interface
120‧‧‧處理器 120‧‧‧ processor
130‧‧‧使用者介面 130‧‧‧user interface
140‧‧‧資料庫 140‧‧‧Database
150a~150n‧‧‧雲環境 150a ~ 150n‧‧‧cloud environment
200‧‧‧雲端資源管理方法 200‧‧‧ Cloud Resource Management Method
300‧‧‧雲端資源管理方法 300‧‧‧ Cloud Resource Management Method
400‧‧‧雲端資源管理方法 400‧‧‧ Cloud resource management method
500‧‧‧雲端資源管理裝置 500‧‧‧ cloud resource management device
510‧‧‧辨識指令模組 510‧‧‧Identification command module
520‧‧‧雲端資源需求項目取得模組 520‧‧‧ Cloud Resource Demand Project Acquisition Module
530‧‧‧資源組合計算與選定模組 530‧‧‧Resource combination calculation and selection module
540‧‧‧機器學習模型 540‧‧‧ Machine Learning Model
550‧‧‧請求發送模組 550‧‧‧Request to send module
S210~S260‧‧‧步驟 S210 ~ S260‧‧‧step
S305~S335‧‧‧步驟 S305 ~ S335‧‧‧step
S410~S440‧‧‧步驟 S410 ~ S440‧‧‧step
以下詳細描述結合隨附圖式閱讀時,將有利於較佳地理解本揭示文件之態樣。應注意,根據說明上實務的需求,圖式中各特徵並不一定按比例繪製。實際上,出於論述清晰之目的,可能任意增加或減小各特徵之尺寸。 The following detailed description, when read in conjunction with the accompanying drawings, will facilitate a better understanding of the aspects of this disclosure. It should be noted that, according to the practical requirements of the description, the features in the drawings are not necessarily drawn to scale. In fact, the size of each feature may be arbitrarily increased or decreased for clarity of discussion.
第1圖繪示根據本揭示文件一些實施例中一種雲端資源管理系統的功能方塊示意圖。 FIG. 1 is a functional block diagram of a cloud resource management system according to some embodiments of the present disclosure.
第2圖繪示根據本揭示文件一些實施例中一種雲端資源管理方法的步驟流程圖。 FIG. 2 is a flowchart of steps in a cloud resource management method according to some embodiments of the present disclosure.
第3圖繪示根據本揭示文件一些實施例中一種雲端資源管理方法的步驟流程圖。 FIG. 3 is a flowchart of steps in a cloud resource management method according to some embodiments of the present disclosure.
第4圖繪示根據本揭示文件一些實施例中一種雲端資源管理方法的步驟流程圖。 FIG. 4 is a flowchart of steps in a cloud resource management method according to some embodiments of the present disclosure.
第5圖繪示根據本揭示文件一些實施例中一種雲端資源管理裝置的功能方塊示意圖。 FIG. 5 is a functional block diagram of a cloud resource management device according to some embodiments of the present disclosure.
請參照第1圖,其繪示根據本揭示文件一些實施例中一種雲端資源管理系統100的功能方塊示意圖。雲端資源管理系統100適配於與複數個雲環境150a、150b、....、150n間的介接。在一些實施例中,雲端資源管理系統100用於計算及分配雲端資源需求的運算資源,以於複數個雲環境中分別所屬之雲端資源當中計算出最適合運算任務的資源,及/或完成雲端資源需求所需的最少成本,及/或滿足一或多個偏好條件等。舉例來說,使用者欲建立電子商務系統時,雲端資源管理系統100會考慮複數個雲環境150a~150n,以於該等雲環境之中決定出符合電子商務系統運作需求的雲端資源。某些實施例中,複數個雲環境可指多雲環境,例如指由不同雲端服務提供者提供的複數個雲環境。某些實施例中,複數個雲環境可指混合雲環境,例如複數個雲環境可包含至少一由雲端佈署之環境及一地端佈署之環境。於此意義下,跨雲亦可指跨地端及雲端之資源佈署。某些實施例中,複數個雲環境可包含多雲環境及混合雲環境。某些實施例中,複數個雲環境包含分別包含不同雲端資源之雲環境。例如,不同雲環境包含不同之操作指令、操作方式、虛擬或實體資源類型等。例如,若以不同雲環境執行同一任務,需要執行特定於個別雲環境之操作。 Please refer to FIG. 1, which illustrates a functional block diagram of a cloud resource management system 100 according to some embodiments of the present disclosure. The cloud resource management system 100 is adapted to interface with a plurality of cloud environments 150a, 150b, ..., 150n. In some embodiments, the cloud resource management system 100 is used for computing and allocating computing resources required for cloud resources, so as to calculate the most suitable computing task resource among the cloud resources respectively belonging to the plurality of cloud environments, and / or complete the cloud The minimum cost required for resource requirements, and / or meeting one or more preferences, etc. For example, when a user wants to establish an e-commerce system, the cloud resource management system 100 will consider a plurality of cloud environments 150a-150n to determine cloud resources that meet the operational requirements of the e-commerce system among these cloud environments. In some embodiments, the plurality of cloud environments may refer to a multi-cloud environment, for example, a plurality of cloud environments provided by different cloud service providers. In some embodiments, the plurality of cloud environments may refer to a hybrid cloud environment. For example, the plurality of cloud environments may include at least one environment deployed by the cloud and one environment deployed locally. In this sense, cross-cloud can also refer to deployment of resources across geographies and clouds. In some embodiments, the plurality of cloud environments may include a multi-cloud environment and a hybrid cloud environment. In some embodiments, the plurality of cloud environments include cloud environments each including different cloud resources. For example, different cloud environments include different operation instructions, operation methods, virtual or physical resource types, and so on. For example, if you perform the same task in different cloud environments, you need to perform operations specific to individual cloud environments.
如第1圖所示,雲端資源管理系統100包含網路介面110以及處理器120。網路介面110作為介接於雲端 資源管理系統100與雲環境150a~150n之間的介面,以傳遞和/或接收指令。為使說明簡潔,以下於特定之例示情況中以三個雲環境150a~150c來說明相關操作。應注意的是,本揭示文件的雲端資源管理系統100並不限於操作於三個雲環境,任何數量的雲環境均屬本揭示文件的範疇。如前述,複數個雲環境可包含多雲、混合雲、地端佈署或以上任二者等。 As shown in FIG. 1, the cloud resource management system 100 includes a network interface 110 and a processor 120. The network interface 110 is connected to the cloud An interface between the resource management system 100 and the cloud environments 150a-150n to transmit and / or receive instructions. In order to make the description concise, the following describes the related operations with three cloud environments 150a-150c in a specific example case. It should be noted that the cloud resource management system 100 of the present disclosure is not limited to operating in three cloud environments, and any number of cloud environments are within the scope of the present disclosure. As mentioned above, the plurality of cloud environments may include multi-cloud, hybrid cloud, ground deployment, or any of the above.
複數個雲環境如雲環境150a~150c分別包含複數個雲端資源。雲端資源可包括但不限於虛擬及/或實體資源。例如,虛擬資源可包括虛擬化資源、軟體定義資源等。舉例而言,虛擬資源可包括一或多個虛擬處理器或虛擬機。又如虛擬資源可包括以不同實體資源為基礎所定義之資源,例如虛擬資源可包括以不同實體伺服器為基礎之一或多個虛擬私雲。又如虛擬資源可包括跨地理位置之資源。又如虛擬資源可包括跨複數個雲環境之雲端資源。在一些實施例中,複數個雲環境如雲環境150a~150c中的雲端資源為包含規格全部相同/等效或規格部分相同/等效的資源。等效規格可包含效能相同、同等級、或皆能滿足特定需求之規格。等效規格可指對單一類型之複數個規格之間的效能比對,抑或對包含複數個類型之複數個規格總成之間的效能比對。例如,一第一規格可包含四個虛擬處理器及對應各虛擬處理器的一個16GB記憶體,而與第一規格等效之一第二規格可包含二個虛擬處理器及對應各虛擬處理器的一個32GB記 憶體。於某些實施例中,第一規格的四個虛擬處理器可等效於第二規格的二個虛擬處理器,及/或第一規格的四個記憶體可等效於第二規格的二個記憶體,及/或第一規格等效於第二規格。於某些實施例中,第一規格與的二規格等效係指第一規格與第二規格的效能相同、或具相同等級、或皆能滿足某一雲端服務的需求。例如,等效之第一規格與第二規格可包含相同、或具相同等級、或皆能滿足特定雲端服務需求的運算能力。又如,一個雲環境之一雲端儲存資源可等效於另外二個雲環境之複數個雲端儲存資源。某些實施例中,雲端資源可以各類型資訊描述。例如,描述雲端資源之資訊可以包含資源種類資訊、資源規格資訊、資源數量資訊、資源時間資訊、資源價格資訊、收費方式資訊、資源狀態資訊等。資源種類資訊可包括描述虛擬機、容器、資料庫、負載器、閘道等不同雲端資源之資訊。資源規格資訊可包括描述資源規格、資源版本、資源型號、資源等級等資訊。資源規格可包括相關於處理器、記憶體、儲存空間、每秒輸出輸入(IOPS)、傳輸量(throughput)、頻寬等資訊。在某些實施例中,作業系統或其版本之相關資訊可作為資源種類資訊或資源規格資訊。作業系統可包括Windows、Linux、或其他雲端適用之作業系統。資源數量資訊可包括描述可用資源數量、最大可用資源數量、最小可用資源數量、不固定數量、隨需數量等資訊。資源時間資訊可包括描述啟動資源之時間、暫停資源之時 間、釋放資源之時間、重啟資源之時間、資源可用期間、資源生命週期、資源在線/離線時間或期間、資源競標時間等資訊。資源價格資訊可包括隨需價格、固定或彈性價格、備用或預留資源價格、資料傳輸價格、競標價格等資訊。收費方式資訊可包括預付款、不預付款、分期付款等資訊。資源狀態資訊可包含資源使用或可用狀態、資源使用或可用條件等資訊。在某些實施例中,前述各類資訊的內容可依各雲環境之所屬分類而與前述分類不同。 The plurality of cloud environments such as the cloud environments 150a to 150c each include a plurality of cloud resources. Cloud resources may include, but are not limited to, virtual and / or physical resources. For example, virtual resources may include virtualization resources, software-defined resources, and the like. For example, the virtual resources may include one or more virtual processors or virtual machines. For another example, the virtual resources may include resources defined based on different physical resources. For example, the virtual resources may include one or more virtual private clouds based on different physical servers. Another example is that virtual resources may include resources across geographic locations. For another example, the virtual resources may include cloud resources across multiple cloud environments. In some embodiments, the cloud resources in the plurality of cloud environments, such as cloud environments 150a-150c, are resources that include all the same / equivalent specifications or the same / equivalent specifications. Equivalent specifications can include specifications that have the same performance, the same level, or can meet specific needs. The equivalent specification may refer to a performance comparison between a plurality of specifications of a single type, or a performance comparison between a plurality of specification assemblies including a plurality of types. For example, a first specification may include four virtual processors and a 16GB memory corresponding to each virtual processor, and a second specification equivalent to the first specification may include two virtual processors and corresponding virtual processors. A 32GB note Memory. In some embodiments, the four virtual processors of the first specification may be equivalent to two virtual processors of the second specification, and / or the four memories of the first specification may be equivalent to two of the second specifications of the second processor. Memory, and / or the first specification is equivalent to the second specification. In some embodiments, the equivalent of the first specification and the second specification means that the performance of the first specification and the second specification are the same, or have the same level, or both can meet the requirements of a certain cloud service. For example, the equivalent first specification and the second specification may include the same, or the same level, or both can meet the computing capabilities of specific cloud service requirements. For another example, one cloud storage resource in one cloud environment may be equivalent to multiple cloud storage resources in the other two cloud environments. In some embodiments, cloud resources can be described by various types of information. For example, the information describing cloud resources may include resource type information, resource specification information, resource quantity information, resource time information, resource price information, charging method information, resource status information, and so on. Resource type information can include information describing different cloud resources such as virtual machines, containers, databases, loaders, and gateways. Resource specification information may include information describing resource specifications, resource versions, resource models, resource levels, and so on. The resource specifications may include information related to the processor, memory, storage space, IOPS, throughput, bandwidth, and so on. In some embodiments, the relevant information of the operating system or its version can be used as resource type information or resource specification information. The operating system may include Windows, Linux, or other cloud-based operating systems. The resource quantity information may include information describing available resource quantity, maximum available resource quantity, minimum available resource quantity, non-fixed quantity, and on-demand quantity. Resource time information may include a description of when the resource was started, when it was suspended Time, resource release time, resource restart time, resource availability period, resource life cycle, resource online / offline time or period, resource bidding time and other information. Resource price information can include on-demand prices, fixed or flexible prices, spare or reserved resource prices, data transfer prices, bid prices, and more. The charging method information may include information such as prepayment, non-prepayment, and installment payment. Resource status information can include information about resource use or availability status, resource use or availability conditions, and more. In some embodiments, the content of the foregoing types of information may be different from the foregoing classification according to the classification of each cloud environment.
如第1圖所示,雲端資源管理系統100包含網路介面110以及處理器120。網路介面110作為在雲端資源管理系統100與雲環境150a~150n之間直接或接間的介面,以傳遞和/或接收指令、資料等。例如,處理器120可透過網路介面110對雲環境150a~150n進行應用程式介面(API)呼叫。 As shown in FIG. 1, the cloud resource management system 100 includes a network interface 110 and a processor 120. The network interface 110 serves as an interface directly or indirectly between the cloud resource management system 100 and the cloud environments 150a to 150n to transmit and / or receive instructions, data, and the like. For example, the processor 120 may make an application program interface (API) call to the cloud environments 150a-150n through the network interface 110.
在一些實施例中,雲端資源管理系統100的處理器120用以依據一雲端資源需求來向雲環境150a~150n中的雲端資源中的一或多個請求配置,以執行能滿足該雲端資源需求的任務。例如,根據一使用者之需求藉由雲端資源管理系統100來配置雲環境150a~150n中的一或多個雲端資源。某些實施例中,一雲端資源需求係由雲端資源管理系統100自動產生、因應事件產生、或週期性產生而使雲端資源管理系統100執行雲端資源配置之任務。例如,雲端資源管理系統100 的處理器120可不直接從使用者接獲指令而執行本文件中所揭露之各項任務。 In some embodiments, the processor 120 of the cloud resource management system 100 is configured to request one or more of the cloud resources in the cloud environments 150a to 150n according to a cloud resource demand, so as to execute a request that can meet the cloud resource demand. task. For example, one or more cloud resources in the cloud environments 150a-150n are configured by the cloud resource management system 100 according to a user's needs. In some embodiments, a cloud resource requirement is generated by the cloud resource management system 100 automatically, in response to an event, or periodically, causing the cloud resource management system 100 to perform the task of cloud resource allocation. For example, cloud resource management system 100 The processor 120 may perform various tasks disclosed in this document without directly receiving instructions from the user.
為易於說明如何請求分配雲端資源的配置及操作,請一併參照第1圖及第2圖。第2圖繪示根據本揭示文件一些實施例中一種雲端資源管理方法200的步驟流程圖。雲端資源管理系統100執行雲端資源管理方法200的各步驟,以佈建最適合的雲端資源。 For easy explanation of how to request the allocation and operation of cloud resource allocation, please refer to Figure 1 and Figure 2 together. FIG. 2 is a flowchart of steps in a cloud resource management method 200 according to some embodiments of the present disclosure. The cloud resource management system 100 executes the steps of the cloud resource management method 200 to deploy the most suitable cloud resources.
在步驟S210中,透過處理器120辨識一指令,其中指令定義一雲端資源需求。 In step S210, an instruction is identified through the processor 120, where the instruction defines a cloud resource requirement.
某些實施例中,雲端資源管理系統100的處理器120可自使用者接收一指令,或接收自動產生、因應事件產生、或週期性產生之一指令。指令可包括一雲端資源需求之資訊。例如,雲端資源需求可包含於某時間啟動特定數量之虛擬機(及/或其他雲端資源)。該時間點可被定義為絕對時間(如2019年6月5日上午10點鐘)、相對時間(如15個小時以後)、由某特定事件觸發之時間點(如虛擬機或處理器使用量達80%)、特定期間(如2019年6月1日至2019年9月30日,或每周二晚上10點鐘)中至少一者。雲端資源需求可包括不限定於特定雲環境或特定雲端資源之需求。例如,雲端資源需求僅定義特定運算能力或特定啟動雲端資源之時間,惟不限定以哪個雲環境或何種雲端資源或資源數量滿足需求。而指令之自動、因應特定事件或週期性之產生亦可如前述對時間點定義之各方式產生。又舉例來說,當使用者欲建立擁有 電子商務系統的執行能力但又不想要自行購買硬體設備時,則透過例如網路通訊協定、使用者介面130或API等可與雲端資源管理系統100通訊的方式,傳送指令至雲端資源管理系統100。某些實施例中,處理器120讀取指令之後,可對指令進行如轉譯或轉換等處理,解讀雲端資源需求或擷取到雲端資源需求,以執行後續的雲端資源佈建。例如,處理器120處理指令後可產生之一或多個後續指令,並將該一或多個後續指令分別傳送至一或多個複數個雲環境。 In some embodiments, the processor 120 of the cloud resource management system 100 may receive an instruction from a user, or receive an instruction that is automatically generated, generated in response to an event, or generated periodically. The instruction may include information on a cloud resource requirement. For example, cloud resource requirements may include starting a specific number of virtual machines (and / or other cloud resources) at a certain time. The time point can be defined as absolute time (such as 10 am on June 5, 2019), relative time (such as 15 hours later), and the time point triggered by a specific event (such as virtual machine or processor usage) Up to 80%), at least one of a certain period (such as June 1, 2019 to September 30, 2019, or every Tuesday at 10 o'clock in the evening). Cloud resource requirements may include requirements that are not limited to specific cloud environments or specific cloud resources. For example, cloud resource requirements only define specific computing power or specific time to start cloud resources, but not limited to which cloud environment or cloud resource or amount of resources to meet the demand. And the generation of instructions automatically, in response to a specific event, or periodically can also be generated in various ways as defined above for time points. As another example, when a user wants to create When the e-commerce system has the execution capability but does not want to purchase hardware equipment by itself, it can send instructions to the cloud resource management system through a method such as a network communication protocol, user interface 130, or API that can communicate with the cloud resource management system 100 100. In some embodiments, after the processor 120 reads the instructions, it can perform processing such as translation or conversion on the instructions, interpret cloud resource requirements or retrieve cloud resource requirements to perform subsequent cloud resource deployment. For example, the processor 120 may generate one or more subsequent instructions after processing the instructions, and transmit the one or more subsequent instructions to one or more cloud environments, respectively.
在步驟S220中,經由網路介面110自複數個雲環境如雲環境150a~150c獲得關聯於指令的雲端資源需求之複數個雲端資源需求項目。 In step S220, a plurality of cloud resource demand items related to the commanded cloud resource demand are obtained from a plurality of cloud environments such as cloud environments 150a-150c via the network interface 110.
在一些實施例中,雲端資源需求項目包含雲端資源規格。雲端資源需求項目可以是資源種類、資源規格、資源數量、資源時間、資源價格、收費方式等。雲端資源需求項目可包含資源種類資訊、資源規格資訊、資源數量資訊、資源時間資訊、資源價格資訊、收費方式資訊、資源狀態資訊等資訊。雲端資源需求項目可包含參數值,參數值用以指出雲端資源需求項目的所需要資源量和/或運算能力。某些實施例中,參數值可對應前述雲端資源需求項目所包含之資訊。在一些實施例中,參數值可包含資源種類資訊、資源規格資訊、資源數量資訊、資源時間資訊、資源價格資訊、收費方式資訊、資源狀態資訊中至少一者。某些實施例中,處理器 120經由網路介面110向複數個雲環境發出請求以獲取關聯於指令的雲端資源需求之複數個雲端資源需求項目。處理器120可於解讀或擷取指令內所載之需求內容之後,向複數個雲環境發送回傳雲端資源需求項目的請求。雲端資源管理系統100可接收並儲存複數個雲環境所反饋的雲端資源需求項目於一儲存器150,以供後續操作。某些實施例中,處理器120可分析指令中之內容,並可檢查複數個雲環境所反饋的雲端資源需求項目是否能充分反映指令的內容。例如,若指令的需求內容包含於某特定時間啟動特定雲端資源,則處理器120可檢查複數個雲環境所反饋的雲端資源需求項目是否包含該特定時間及該特定雲端資源之資訊。某些實施例中,若特定雲環境所反饋的雲端資源需求項目未包含該特定時間及該特定雲端資源之資訊,處理器120可將該特定雲環境從複數個雲環境選項中剃除,或於本次任務中不再考慮該特定雲環境。某些實施例中,雲端資源管理系統100可週期性或依事件觸發等方式接收複數個雲環境的雲端資源需求項目。例如,本次任務中可依據之雲端資源需求項目可為最近一次接收或最新之雲端資源需求項目。 In some embodiments, the cloud resource requirement item includes a cloud resource specification. Cloud resource demand items can be resource types, resource specifications, resource quantities, resource time, resource prices, charging methods, etc. Cloud resource requirement items can include information such as resource type information, resource specification information, resource quantity information, resource time information, resource price information, charging method information, resource status information, and so on. The cloud resource requirement project may include parameter values, and the parameter value is used to indicate the required resource amount and / or computing power of the cloud resource requirement project. In some embodiments, the parameter value may correspond to the information contained in the aforementioned cloud resource requirement item. In some embodiments, the parameter value may include at least one of resource type information, resource specification information, resource quantity information, resource time information, resource price information, charging method information, and resource status information. In some embodiments, the processor 120 sends a request to a plurality of cloud environments via the network interface 110 to obtain a plurality of cloud resource demand items associated with the commanded cloud resource demand. The processor 120 may send a request for returning a cloud resource requirement item to a plurality of cloud environments after interpreting or retrieving the demand content contained in the instruction. The cloud resource management system 100 may receive and store cloud resource demand items fed back from a plurality of cloud environments in a storage 150 for subsequent operations. In some embodiments, the processor 120 may analyze the content in the instruction, and may check whether the cloud resource requirement items fed back by the plurality of cloud environments can fully reflect the content of the instruction. For example, if the requested content of the instruction includes starting a specific cloud resource at a specific time, the processor 120 may check whether the cloud resource demand items fed back from the plurality of cloud environments include information of the specific time and the specific cloud resource. In some embodiments, if the cloud resource requirement item fed back by the specific cloud environment does not include the information of the specific time and the specific cloud resource, the processor 120 may remove the specific cloud environment from a plurality of cloud environment options, or This particular cloud environment is no longer considered in this task. In some embodiments, the cloud resource management system 100 may receive a plurality of cloud resource demand items of the cloud environment periodically or in an event triggered manner. For example, the cloud resource demand items that can be based on this task can be the most recently received or the latest cloud resource demand items.
在一些實施例中,處理器120獲得的雲端資源項目中的資源種類資訊與資源數量資訊的參數值例如是4個處理器、16GB的記憶體、32GB的儲存資源、10GB的輸出輸入等資訊。處理器120獲得的雲端資源項 目中的資源價格資訊參數值例如是啟用一個處理器需要花費每小時$0.4328等資訊。 In some embodiments, the parameter values of the resource type information and the resource quantity information in the cloud resource item obtained by the processor 120 are, for example, information of 4 processors, 16GB of memory, 32GB of storage resources, 10GB of input and output. Cloud resource items obtained by the processor 120 The current resource price information parameter values are information such as enabling a processor to cost $ 0.4328 per hour.
在步驟S230中,藉由處理器120根據指令針對所針對雲端資源需求項目的參數值計算出複數個資源組合。 In step S230, the processor 120 calculates a plurality of resource combinations for the parameter values of the target cloud resource demand item according to the instruction.
舉例來說,處理器120於前述步驟S220中取得雲端資源需求項目的參數值包含12個處理器以及128GB記憶體,處理器120可判斷複數個雲環境如雲環境150a~150c的複數個雲端資源當中滿足或符合此參數值的複數個資源組合。 For example, the parameter value of the cloud resource requirement item obtained by the processor 120 in the foregoing step S220 includes 12 processors and 128GB of memory. The processor 120 may determine a plurality of cloud resources, such as a plurality of cloud resources in a cloud environment 150a to 150c Multiple resource combinations that meet or meet the value of this parameter.
請參照下表一至表三,分別為雲環境150a~150c的雲端資源實施例。 Please refer to the following Tables 1 to 3 for the cloud resource embodiments of the cloud environment 150a ~ 150c, respectively.
如上表一至表三所例示,雲端資源需求項目可包括處理器、記憶體、儲存資源、資源價格、作業系統中至少複數個需求項目,而依需求項目不同的參數值(或參數值的組合)而有不同的配置。某些實施例中,處理器120可根據雲端資源需求項目的參數值,計算出表一至表三的雲端資源列表的複數個資源組合。某些實施例中,雲端資源管理系統100可正規化所反饋之雲端資源需求項目的參數值。例如,不同雲環境可能對同一需求項目有不同的定義或設定而造成專屬特定雲環境之參數值,而雲端資源管理系統100可將此種基於雲環境 特徵的差異予以消彌,以獲得基於同一標準之正規化參數值。 As exemplified in Tables 1 to 3 above, cloud resource demand items can include processors, memory, storage resources, resource prices, at least a plurality of demand items in the operating system, and different parameter values (or a combination of parameter values) according to the demand items. There are different configurations. In some embodiments, the processor 120 may calculate a plurality of resource combinations of the cloud resource list of Tables 1 to 3 according to parameter values of the cloud resource demand items. In some embodiments, the cloud resource management system 100 may normalize the parameter values of the cloud resource demand items that are fed back. For example, different cloud environments may have different definitions or settings for the same demand item, resulting in parameter values specific to a specific cloud environment, and the cloud resource management system 100 may apply such cloud-based environments Differences in features are eliminated to obtain normalized parameter values based on the same criteria.
在某些實施例中,若存在一雲端運算需求要在例如中午12點時,啟動12個處理器以及128GB記憶體,則處理器120取得對應雲環境150a~150c的相關規格,以產生複數個資源組合,如產生所針對雲環境中所有可能的資源組合。請參照下表四,表四為複數個資源組合的實施例。 In some embodiments, if there is a cloud computing requirement to start 12 processors and 128GB of memory at 12 noon, the processor 120 obtains the relevant specifications corresponding to the cloud environment 150a ~ 150c to generate a plurality of Resource combinations, such as all possible resource combinations in the targeted cloud environment. Please refer to Table 4 below, which is an example of a plurality of resource combinations.
在需要12個處理器以及128GB記憶體的情況,處理器120計算出四個組合,惟本創作不限於此。第一組合包含雲環境150a的第三配置,需求共16單位,即16組雲環境150a的第三配置(下稱第一資源組合)。第 二組合包含雲環境150a的第五配置,需求共8單位,即8組雲環境150a的第五配置(下稱第二資源組合)。第三組合包含雲環境150a的第三配置,需求共4單位(即4組雲環境150a的第三配置),雲環境150b的第二配置,需求共4單位(即4組雲環境150b的第二配置),以及雲環境150c的第五配置,需求共1單位(即1組雲環境150c的第五配置)(下稱第三資源組合)。第四組合包含雲環境150b的第四配置,需求共1單位(即4組雲環境150b的第四配置),以及雲環境150c的第四配置,需求共2單位(即2組雲環境150c的第四配置)(下稱第四資源組合)。某些實施例中,如表四列舉之各個資源組合均為包含等效規格的資源組合。例如,此等資源組合兩兩之間均為等效,即各資源組合均能產生相同效能或符合該雲端運算需求。有關等效之實施例可參照本文件其他相關部分。 When 12 processors and 128GB of memory are required, the processor 120 calculates four combinations, but this creation is not limited to this. The first combination includes the third configuration of the cloud environment 150a, which requires a total of 16 units, that is, the third configuration of the 16 groups of cloud environments 150a (hereinafter referred to as the first resource combination). First The second combination includes the fifth configuration of the cloud environment 150a, which requires a total of 8 units, that is, the fifth configuration of the eight groups of cloud environments 150a (hereinafter referred to as the second resource combination). The third combination includes the third configuration of the cloud environment 150a, which requires a total of 4 units (that is, the third configuration of four groups of cloud environments 150a), and the second configuration of the cloud environment 150b, which requires a total of 4 units (that is, the first group of 4 groups of cloud environments 150b) Two configurations), and the fifth configuration of the cloud environment 150c, requires a total of 1 unit (that is, the fifth configuration of a group of cloud environments 150c) (hereinafter referred to as the third resource combination). The fourth combination includes the fourth configuration of cloud environment 150b, which requires a total of 1 unit (that is, the fourth configuration of 4 groups of cloud environments 150b), and the fourth configuration of cloud environment 150c, which requires a total of 2 units (that is, 2 groups of cloud environments 150c Fourth configuration) (hereinafter referred to as the fourth resource combination). In some embodiments, each resource combination listed in Table 4 is a resource combination including an equivalent specification. For example, these resource combinations are equivalent to each other, that is, each resource combination can produce the same performance or meet the cloud computing requirements. For equivalent embodiments, refer to other relevant parts of this document.
在步驟S240中,藉由處理器120自該些資源組合中選定一資源組合,其中被選定的資源組合可滿足指令之雲端資源配置需求。 In step S240, the processor 120 selects a resource combination from the resource combinations, and the selected resource combination can meet the commanded cloud resource allocation requirements.
舉例來說,處理器120根據雲端資源需求項目的參數值,計算出如上表四的第一資源組合至第四資源組合。 For example, the processor 120 calculates the first resource combination to the fourth resource combination as shown in Table 4 above according to the parameter values of the cloud resource demand items.
在一些實施例中,處理器120執行數值演算法、機器學習演算法或其他演算法,以計算出複數個資源組合。舉例來說,處理器120使用線性規劃(Linear Programming)或非線性規劃(Non-Linear Programming),根據資源組合的參數值來計算組合需求值。或者,處理器120使用機器學習回歸分析,例如最小方均根、隨機森林回歸因子(Regress Forest Regressor)、線性回歸(Linear Regression)、支持向量機(Support Vector Regression,SVR)等方式,計算出不同之資源組合。 In some embodiments, the processor 120 executes a numerical algorithm, a machine learning algorithm, or other algorithms to calculate a plurality of resource combinations. For example, the processor 120 uses Linear Programming or Non-Linear Programming. Programming), to calculate the combined demand value according to the parameter value of the resource combination. Alternatively, the processor 120 uses machine learning regression analysis, such as minimum root mean square, random forest regression factor (Regress Forest Regressor), linear regression (Linear Regression), support vector machine (Support Vector Regression (SVR), etc.) to calculate different resources combination.
處理器120可獲得第一資源組合、第二資源組合、第三資源組合以及第四資源組合後,選擇其一來作為選定的資源組合,以滿足指令之雲端資源配置需求。某些實施例中,可能有複數個資源組合皆能滿足雲端資源配置需求時,處理器120可進一步判斷是否其他條件來輔助其選擇其中一資源組合。例如,處理器120可根據一選取規則選擇滿足雲端資源配置需求之複數個資源組合中總價格最低者;或者,處理器120可據以選擇滿足雲端資源配置需求之複數個資源組合中平均效能參數值最高者。某些實施例中,選取規則可依雲端資源管理政策設定或調整。例如,雲端資源管理政策可反映出使用者偏好,使用者可設定當前雲端資源管理政策為成本最小導向,故在此政策下,處理器120可依相應規則選取價格最低之資源組合;或者,雲端資源管理政策可設為虛擬機數量最大化並配合處理器本身效能最小化之導向,使處理器120較偏向選取低階處理機之資源組合。 After the processor 120 obtains the first resource combination, the second resource combination, the third resource combination, and the fourth resource combination, the processor 120 selects one of them as the selected resource combination to meet the commanded cloud resource allocation requirements. In some embodiments, when there may be multiple resource combinations that can meet the cloud resource configuration requirements, the processor 120 may further determine whether other conditions assist it in selecting one of the resource combinations. For example, the processor 120 may select the lowest total price among a plurality of resource combinations that satisfy the cloud resource allocation requirements according to a selection rule; or, the processor 120 may select an average performance parameter among the plurality of resource combinations that meet the cloud resource allocation requirements. The highest value. In some embodiments, the selection rule may be set or adjusted according to a cloud resource management policy. For example, the cloud resource management policy can reflect user preferences. The user can set the current cloud resource management policy to be the least cost-oriented. Therefore, under this policy, the processor 120 can select the lowest-priced resource combination according to the corresponding rules; or, the cloud The resource management policy can be set to maximize the number of virtual machines and to minimize the performance of the processor itself, so that the processor 120 prefers to select a low-level processor resource combination.
在一些實施例中,雲端資源需求包含目標需 求值,例如資源成本。處理器120會根據各資源組合的參數值來計算組合需求值,以於複數個資源組合中選定其一。組合需求值例如是資源成本。舉例來說,第一資源組合的資源成本是$23.8512,第二資源組合的資源成本是$43.7592,第三資源組合的資源成本是$12.4407,第四資源組合的資源成本是$8.338。在某些實施例中,雲端資源需求可包含對特定參數值如某資源價格的限定,則處理器120可根據該限定而先排除不符合的資源配置。例如,若使用者要求排除價格高於每小時$4的配置,則處理器120可先剔除每小時$4的所有配置,再行計算資源組合。又如,若欲僅使用最新版本之資源,則處理器120可先剔除所有舊版本之配置,再行計算資源組合。據此,在某些情況下,可大幅減少計算量。 In some embodiments, cloud resource requirements include target requirements Evaluation, such as resource costs. The processor 120 calculates a combination demand value according to the parameter value of each resource combination, so as to select one of the plurality of resource combinations. The combined demand value is, for example, a resource cost. For example, the resource cost of the first resource combination is $ 23.8512, the resource cost of the second resource combination is $ 43.7592, the resource cost of the third resource combination is $ 12.4407, and the resource cost of the fourth resource combination is $ 8.338. In some embodiments, the cloud resource requirements may include restrictions on specific parameter values, such as a resource price, and the processor 120 may first exclude non-compliant resource configurations according to the restrictions. For example, if the user requests to exclude configurations with a price higher than $ 4 per hour, the processor 120 may remove all configurations of $ 4 per hour before calculating the resource combination. For another example, if only the latest version of the resources is to be used, the processor 120 may first remove all the configurations of the old version before calculating the resource combination. According to this, in some cases, the calculation amount can be greatly reduced.
處理器120根據目標需求值與各資源組合的組合需求值來選定一資源組合,例如比對目標需求值與各資源組合的組合需求值而選定其中一資源組合。 The processor 120 selects a resource combination according to the target demand value and the combined demand value of each resource combination, for example, compares the target demand value and the combined demand value of each resource combination to select one of the resource combinations.
在一些實施例中,處理器120於組合需求值中小於目標需求值的資源組合來作為選定的資源組合。舉例來說,若目標需求值是指$30.0的資源成本,則處理器120選擇資源成本低於$30.0的資源組合,如資源成本為$12.4407的第三資源組合。 In some embodiments, the processor 120 selects a resource combination that is smaller than the target demand value in the combined demand value as the selected resource combination. For example, if the target demand value is a resource cost of $ 30.0, the processor 120 selects a resource combination with a resource cost less than $ 30.0, such as a third resource combination with a resource cost of $ 12.4407.
在另一些實施例中,處理器120根據資源組合的組合需求值中的最小者來選擇資源組合。舉例來說,處理器120會選擇資源成本最小的第四資源組合。 In other embodiments, the processor 120 selects a resource combination according to the smallest of the combined demand values of the resource combination. For example, the processor 120 may select a fourth resource combination with the lowest resource cost.
在步驟S250中,經由網路介面110向第一雲環境送出第一請求,以於第一雲環境中配置第一雲端資源,其中第一雲端資源屬於該些雲環境中之第一雲環境。 In step S250, a first request is sent to the first cloud environment via the network interface 110 to configure the first cloud resource in the first cloud environment, where the first cloud resource belongs to the first cloud environment in the cloud environments.
舉例來說,當處理器120選擇表四中之第四資源組合,如前述步驟S240所述,第四資源組合包含雲環境150b的雲端資源之1單位的第四配置。因此,經由網路介面110向雲環境150b送出請求,以於雲環境150b中配置1單位的16個處理器、1單位的32GB記憶體以及CentOS作業系統。雲環境150b可依據該請求而進行相應配置。 For example, when the processor 120 selects the fourth resource combination in Table 4, as described in the foregoing step S240, the fourth resource combination includes a fourth configuration of one unit of cloud resources of the cloud environment 150b. Therefore, a request is sent to the cloud environment 150b via the network interface 110, so that one unit of 16 processors, one unit of 32GB memory, and a CentOS operating system are configured in the cloud environment 150b. The cloud environment 150b can be configured accordingly according to the request.
在步驟S260中,經由網路介面110向第二雲環境送出第二請求,以於第二雲環境中配置第二雲端資源,其中第二雲端資源屬於該些雲環境中之第二雲環境。 In step S260, a second request is sent to the second cloud environment via the network interface 110 to configure a second cloud resource in the second cloud environment, where the second cloud resource belongs to the second cloud environment in the cloud environments.
舉例來說,當處理器120選擇第四資源組合,如前述步驟S240所述,第四資源組合還包含雲環境150c的雲端資源之2單位的第四配置。因此,經由網路介面110向雲環境150c送出請求,以於雲環境150c中配置2單位的8個處理器以及2單位的30GB記憶體以及2單位的16GB儲存資源。雲環境150c可依據該請求而進行相應配置。據此,雲端資源管理系統100針對指令可依本文件所揭露之方式將需求分配給雲環境150b及雲環境150c,如此使用者不需煩惱雲環境的選擇及配置細節而仍可 完成需求。 For example, when the processor 120 selects a fourth resource combination, as described in step S240, the fourth resource combination further includes a fourth configuration of two units of cloud resources of the cloud environment 150c. Therefore, a request is sent to the cloud environment 150c via the network interface 110, so that two units of eight processors, two units of 30GB memory, and two units of 16GB storage resources are configured in the cloud environment 150c. The cloud environment 150c can be configured accordingly according to the request. According to this, the cloud resource management system 100 can allocate requirements to the cloud environment 150b and cloud environment 150c according to the instructions disclosed in this document, so that users can still use the cloud environment without having to worry about the selection and configuration details of the cloud environment. Complete the requirements.
在某些實施例中,處理器120可辨識關聯於雲端資源需求項目至少一者之一權重值,以作為計算資源組合之依據。在一些實施例中,經由網路介面110於複數個雲環境如雲環境150a~150c獲得關聯於步驟S210所讀取到的指令時,處理器120根據指令辨識關聯於雲端資源需求項目的權重值。接著,藉由處理器120根據雲端資源需求項目的權重值與雲端資源需求項目的參數值,來計算出資源組合。值得一提的是,雲端資源需求項目的權重值可以為選擇性的參數,此權重值用來提高或降低雲端資源項目被選擇的機率。舉例來說,若上表二的作業系統(或其發行版本)被設定有權重值,例如Red Hat的權重值為0.8以及CentOS的權重值為0.4,若指令中也包含有作業系統的權重值之要求,則處理器120會以權重值高的雲端資源作為優先考量,例如Red Hat的作業系統相較於CentOS會被優先選擇。 In some embodiments, the processor 120 may identify a weight value associated with at least one of the cloud resource demand items as a basis for calculating a resource combination. In some embodiments, when the instructions associated with the read in step S210 are obtained from a plurality of cloud environments such as cloud environments 150a to 150c via the network interface 110, the processor 120 identifies the weight value associated with the cloud resource requirement item according to the instruction . Then, the processor 120 calculates a resource combination according to the weight value of the cloud resource demand item and the parameter value of the cloud resource demand item. It is worth mentioning that the weight value of the cloud resource demand item can be an optional parameter, and this weight value is used to increase or decrease the probability that the cloud resource item is selected. For example, if the operating system (or its release version) in Table 2 is set with a weight value, for example, the weight value of Red Hat is 0.8 and the weight value of CentOS is 0.4. If the command also includes the weight value of the operating system If required, the processor 120 will prioritize cloud resources with a high weight value. For example, Red Hat's operating system will be preferred over CentOS.
在一些實施例中,雲端資源需求項目的權重值可根據被選擇的次數,來增加雲端資源需求項目的權重值。舉例來說,雲環境150b的雲端資源之第四配置與雲環境150c的雲端資源之第四配置被選擇,因此在該次的資源配置後,雲環境150b的雲端資源之第四配置與雲環境150c的雲端資源之第四配置的選擇或配置記錄將作為歷史資料,而依此歷史資料更新雲端資源的權重值,例如增加該等配置的權重值。如此一來,雲端資源管理 系統100可學習每一次的資源配置,避免每一次的資源組合或者總資源成本的變化程度太大,以符合使用者的使用習慣。例如,雲端資源管理系統100更選用地包含一資料庫140,用以儲存一包含權重設定規則之模型,以藉由歷史資料訓練模型而更新權重值。在一些實施例中,資料庫140尚可用以儲存歷史資料及/或權重值。 In some embodiments, the weight value of the cloud resource demand item may be increased according to the selected number of times. For example, the fourth configuration of cloud resources in cloud environment 150b and the fourth configuration of cloud resources in cloud environment 150c are selected. Therefore, after this resource configuration, the fourth configuration of cloud resources in cloud environment 150b and cloud environment The selection or configuration record of the fourth configuration of the 150c cloud resource will be used as historical data, and the weight value of the cloud resource is updated according to this historical data, for example, the weight value of such configurations is increased. In this way, cloud resource management The system 100 can learn the resource configuration every time, and avoid the change of each resource combination or the total resource cost to be too large, so as to meet the user's usage habits. For example, the cloud resource management system 100 optionally includes a database 140 for storing a model including weight setting rules to update the weight value by training the model with historical data. In some embodiments, the database 140 may also be used to store historical data and / or weight values.
請復參照第1圖,雲端資源管理系統100包含使用者介面130。使用者介面130耦接於處理器120。使用者介面130提供例如使用者圖形化介面或命令列介面,使得處理器120透過使用者介面130獲得雲端資源需求項目的權重值。如此一來,使用者可透過使用者介面130來設定雲端資源需求項目的權重值。 Please refer to FIG. 1 again, the cloud resource management system 100 includes a user interface 130. The user interface 130 is coupled to the processor 120. The user interface 130 provides, for example, a user graphical interface or a command line interface, so that the processor 120 obtains a weight value of a cloud resource requirement item through the user interface 130. In this way, the user can set the weight value of the cloud resource requirement item through the user interface 130.
如上文提及之雲端資源管理政策,使用者也可以根據實際需求設定其他偏好條件。舉例來說,偏好條件可以是所欲之雲端資源的作業系統(或版本)、數量上限/下限、可用期間、雲端資源的實體存在位置(例如國家或地區或其他地理位置,如機房位置)等。例如,使用者可將雲端資源限制於某地理位置,或將之排除於某地理位置之外。偏好條件亦可以透過權重值設定,例如權重值越高,代表該雲端資源是使用者希望使用的資源。在某些實施例中,偏好條件可進一步將限制條件區分出來,以利增加計算之操作度。例如,偏好條件可包括虛擬機數量、虛擬機數量之上限、喜好之雲環境等,而限制條件可包括作業系統與版本、硬體(如伺服器或處 理器)之新舊與可用期間、雲端服務機房位置等。某些實施例中,偏好條件可作為「軟性」條件,而限制條件可作為「硬性」條件。例如,可賦予軟性條件較低權重值,而賦予硬性條件較高之權重值。 As mentioned in the cloud resource management policy mentioned above, users can also set other preferences based on actual needs. For example, the preference condition can be the operating system (or version) of the desired cloud resource, the upper limit / lower limit, the available period, the physical location of the cloud resource (such as the country or region or other geographic location, such as the location of the computer room), etc. . For example, users can restrict or exclude cloud resources from a certain geographic location. The preference condition can also be set by a weight value. For example, a higher weight value indicates that the cloud resource is a resource that the user wants to use. In some embodiments, the preference condition may further distinguish the restriction condition, so as to increase the operation degree of the calculation. For example, the preferences may include the number of virtual machines, the upper limit of the number of virtual machines, the preferred cloud environment, etc., and the restrictions may include the operating system and version, hardware (such as a server or server) Server), the old and new and available periods, the location of the cloud service room, etc. In some embodiments, a preference condition may be used as a "soft" condition, and a restriction condition may be used as a "hard" condition. For example, a soft condition may be assigned a lower weight value, while a hard condition may be assigned a higher weight value.
在另一些實施例中,處理器120會執行整合移動平均自回歸模型(Autoregressive Integrated Moving Average model,ARIMA),以自動迴歸的樹狀模型將過去的雲端資源使用記錄與所預測的雲端資源需求產生關聯,以估測未來的雲端資源需求。處理器120也可執行遞歸神經網路(Recurrent Neural Network,RNN),使用歷史資料來預測每一個週期的雲端資源需求。如此一來,雲端資源管理系統100不只是依據雲端資源需求來計算所需要的資源組合,還可以藉由歷史資料來預測未來的一段週期內符合雲端資源需求之資源組合。例如,電子商務的流量相當具有規律性,例如平日每天中午的一小時與晚上7點到12點的流量最高,每日凌晨3點到6點流量最低。一週七天在週末時段的流量又比平日高。據此,處理器120可針對當前所使用之雲端資源收集一段時間的流量變化作為歷史資料(如Log),歷史資料可包含雲端資雲之處理器使用狀況(如用量百分比)、記憶體使用狀況、IOPS等資料,可以得到當前所使用之雲端資源的系統效能的歷史曲線。處理器120可將此歷史曲線等作為機器學習的模型訓練值。 In other embodiments, the processor 120 executes an ARIMA (Autoregressive Integrated Moving Average Model) to generate an automatic regression tree model to generate past cloud resource usage records and predicted cloud resource requirements. Correlation to estimate future cloud resource requirements. The processor 120 may also execute a Recurrent Neural Network (RNN) and use historical data to predict cloud resource requirements for each cycle. In this way, the cloud resource management system 100 not only calculates the required resource combinations based on the cloud resource requirements, but also can use historical data to predict the resource combinations that meet the cloud resource requirements in a future period. For example, the e-commerce traffic is quite regular. For example, the traffic at one hour at noon and 7 to 12 pm on weekdays is the highest, and the traffic is lowest from 3 to 6 am. Seven days a week, traffic is higher on weekends than on weekdays. According to this, the processor 120 can collect the traffic changes over a period of time for historically used cloud resources as historical data (such as Log), and the historical data can include the usage status of the cloud-based cloud processor (such as usage percentage), and the memory usage status. Data, IOPS and other data, you can get the historical curve of the system performance of the cloud resources currently used. The processor 120 may use this historical curve or the like as a model training value for machine learning.
舉例來說,如上表一至表三中的雲端資源包 含雲端資源規格。另一方面,處理器120計算得到的資源組合包含等效規格。因此,處理器120根據雲端資源規格來計算與其等效的規格,於所有等效於雲端資源規格的雲端資源當中計算滿足或符合雲端資源需求的雲端資源。舉例來說,本文件所涉及的雲端資源的執行能力符合雲端資源需求者,均可被用於雲端運算當中,並不以硬體及/或虛擬資源完全一樣為限,例如同樣型號、同樣類型、同樣等級、或同樣世代的運算能力的處理器但為不同廠商製造或品牌者,均可屬於等效規格的範疇。 For example, the cloud resource packages in Tables 1 to 3 above Includes cloud resource specifications. On the other hand, the resource combination calculated by the processor 120 includes an equivalent specification. Therefore, the processor 120 calculates the equivalent specifications according to the cloud resource specifications, and calculates the cloud resources that meet or meet the requirements of the cloud resources among all the cloud resources equivalent to the cloud resource specifications. For example, those whose execution capabilities of the cloud resources involved in this document meet the requirements of cloud resources can be used in cloud computing, not limited to the exact same hardware and / or virtual resources, such as the same model and type Processors of the same grade, or the same generation of computing power, but manufactured or branded by different manufacturers, can fall into the category of equivalent specifications.
圖3繪示根據本揭示文件一些實施例中一種雲端資源管理方法300的步驟流程圖。在步驟S305中,雲端資源管理系統100取得一雲端資源需求之所欲總成本及啟動時間。例如,所欲總成本可包含一成本絕對值,或一成本區間。所欲總成本可包含一成本閾值,以作為雲端資源管理系統100於後續計算及/或選擇資源組合之參數。如上文所述,啟動時間可為一絕對時間、週期性時間等。 FIG. 3 is a flowchart of steps in a cloud resource management method 300 according to some embodiments of the present disclosure. In step S305, the cloud resource management system 100 obtains a desired total cost and startup time of a cloud resource demand. For example, the desired total cost may include an absolute cost value, or a cost interval. The desired total cost may include a cost threshold as a parameter for the cloud resource management system 100 to subsequently calculate and / or select a resource combination. As mentioned above, the startup time may be an absolute time, a periodic time, or the like.
在步驟S310中,雲端資源管理系統100取得一或多個偏好條件。該一或多個偏好條件可經由使用者介面130或API自使用者或使用者之系統取得。某些實施例中,偏好條件可已儲存於自雲端資源管理系統100內。偏好條件之實施例可參照如本文件上文相關部分。例如,偏好條件可包括付款條件,如預付款、不預付款、 分期付款等。又如偏好條件可包括雲端資源之地理位置、雲端資源數量、雲端資源價格等。再者,偏好條件可利用如上文所述之雲端資雲管理政策的方式取得。 In step S310, the cloud resource management system 100 obtains one or more preference conditions. The one or more preference conditions can be obtained from the user or the user's system via the user interface 130 or the API. In some embodiments, the preference conditions may have been stored in the self-cloud resource management system 100. For examples of preferred conditions, reference may be made to the relevant sections above in this document. For example, preferences can include payment terms such as prepayment, no prepayment, Installment payments, etc. For another example, the preference conditions may include the geographical location of cloud resources, the number of cloud resources, and the price of cloud resources. Moreover, the preference conditions can be obtained by using the cloud-based cloud management policy as described above.
在步驟S315中,雲端資源管理系統100取得複數個雲環境所提供的複數個雲端資源清單。複數個雲環境可為於通聯方式上不相容之多個雲環境,如各雲環境均有其專屬之API。雲端資源清單可包含如上文所述(並請參照表格一至三)之根據雲端資源需求項目所提供的不同配置。 In step S315, the cloud resource management system 100 obtains a plurality of cloud resource lists provided by a plurality of cloud environments. The plurality of cloud environments may be multiple cloud environments that are incompatible in terms of communication methods. For example, each cloud environment has its own API. The cloud resource list can include different configurations provided by the cloud resource demand project as described above (and please refer to Tables 1 to 3).
在步驟S320中,雲端資源管理系統100將複數個雲端資源清單與偏好條件進行比對,將複數個雲端資源清單中不符合偏好條件的配置剔除,產生計算用之配置。例如,偏好條件將雲端作業系統限制為Linux作業系統時,雲端資源清單中使用非Linux作業系統的配置會被剔除,所剩之配置將可利用做後續的計算。 In step S320, the cloud resource management system 100 compares the plurality of cloud resource lists with the preference conditions, removes the configurations that do not meet the preference conditions from the plurality of cloud resource lists, and generates a configuration for calculation. For example, when the preference condition restricts the cloud operating system to the Linux operating system, the configuration using the non-Linux operating system in the cloud resource list will be eliminated, and the remaining configuration can be used for subsequent calculations.
在步驟S325中,雲端資源管理系統100根據計算用之配置產生多個雲端資源組合,並計算各個雲端資源組合之價格。雲端資源管理系統100可根據本文件揭露之各種方法進行計算。此外,雲端資源管理系統100可賦予計算用之配置中至少一個雲端資源需求項目相應的權重值。有關權重的描述亦可參照本文前揭之實施例。 In step S325, the cloud resource management system 100 generates a plurality of cloud resource combinations according to the calculation configuration, and calculates the price of each cloud resource combination. The cloud resource management system 100 may perform calculations according to various methods disclosed in this document. In addition, the cloud resource management system 100 may assign a corresponding weight value to at least one cloud resource requirement item in the configuration used for computing. For the description of the weights, reference may also be made to the embodiments previously disclosed herein.
在步驟S330中,雲端資源管理系統100根據所欲總成本自多個雲端資源組合中選定一所欲雲端資 源組合。例如,雲端資源管理系統100將多個雲端資源組合的價格與所欲總成本進行比對,於多個雲端資源組合中選取價格低於所欲總成本的雲端資源組合。若有多個價格低於所欲總成本的雲端資源組合,則可選取價格最低者。 In step S330, the cloud resource management system 100 selects a desired cloud resource from a plurality of cloud resource combinations according to a desired total cost. Source combination. For example, the cloud resource management system 100 compares the prices of multiple cloud resource combinations with a desired total cost, and selects a cloud resource combination whose price is lower than the desired total cost among the plurality of cloud resource combinations. If there are multiple cloud resource combinations whose price is lower than the desired total cost, the lowest price can be selected.
在步驟S335中,雲端資源管理系統100根據所選定之所欲雲端資源組合對複數個雲環境進行雲端資源佈署。例如,若所選定之所欲雲端資源組合包含A雲環境及B雲環境的資源時,則依據該雲端資源組合針對A雲環境及B雲環境進行佈署,以滿足雲端資源需求的內容。 In step S335, the cloud resource management system 100 performs cloud resource deployment on a plurality of cloud environments according to the selected desired cloud resource combination. For example, if the selected desired cloud resource combination includes resources of the A cloud environment and the B cloud environment, the A cloud environment and the B cloud environment are deployed according to the cloud resource combination to meet the content of cloud resource requirements.
圖4繪示根據本揭示文件一些實施例中一種雲端資源管理方法400的步驟流程圖。雲端資源管理方法400與雲端資源管理方法300之主要差別在於,於雲端資源管理方法400中,雲端資源需求不包含成本條件及偏好條件,而時間條件則由雲端資源管理系統100可運用之歷史資料來決定。例如,雲端資源管理系統100根據當前(或過去)使用雲端服務記錄之歷史資料,經由機器學習產生未來用量之預測曲線。如上述,雲端資源管理系統100可利用深度學習的遞歸神經網路(RNNs)來預測時間序列,在此情境下,同一個時間單位會有許多類似的特徵時間序列(處理器、IOPS等),而利用不同序列,例如每週、每天、每小時為最小單位的歷史資料,可預測得到系統(雲端服務資源)未來的效能需求曲線。 FIG. 4 is a flowchart of steps in a cloud resource management method 400 according to some embodiments of the present disclosure. The main difference between the cloud resource management method 400 and the cloud resource management method 300 is that in the cloud resource management method 400, cloud resource requirements do not include cost conditions and preference conditions, and time conditions are historical data that can be used by the cloud resource management system 100. To decide. For example, the cloud resource management system 100 generates a forecast curve of future usage through machine learning based on historical data recorded using cloud services currently (or in the past). As described above, the cloud resource management system 100 can use deep learning recurrent neural networks (RNNs) to predict time series. In this scenario, there will be many similar characteristic time series (processors, IOPS, etc.) in the same time unit. And using different sequences, such as weekly, daily, and hourly historical data as the smallest unit, can predict the future performance demand curve of the system (cloud service resources).
在步驟S410中,雲端資源管理方法400辨識一任務需求,該任務需求包含執行一第一任務。例如,任務需求可包含指示繼續執行當前或執行中之一任務,而當前任務已在利用特定之雲端資源配置。或者,任務需求可包含基於一使用特定雲端資源配置而執行完畢之任務,針對該執行完畢之任務於過去的雲端資源配置使用記錄而指示於未來特定時間執行與該執行完畢任務相同之任務。 In step S410, the cloud resource management method 400 identifies a task requirement that includes performing a first task. For example, the task requirements may include instructions to continue performing one of the current or one of the tasks, and the current task is already utilizing a specific cloud resource configuration. Alternatively, the task requirements may include a task that has been completed based on the use of a specific cloud resource configuration, and instructs to perform the same task as the completed task at a specific time in the future for the completed task based on past cloud resource configuration usage records.
在步驟S420中,辨識一第二任務,取得第二任務之歷史資料。如上述,第二任務可為當前執行中或已執行完畢之任務。如本文件所揭露,歷史資料可包含雲端資源配置及/或使用記錄。 In step S420, a second task is identified, and historical data of the second task is obtained. As described above, the second task may be a task that is currently being executed or has been completed. As disclosed in this document, historical data may include cloud resource allocation and / or usage records.
在步驟S430中,利用該歷史資料預測第一任務之雲端資源需求。如本文件所揭露,可透過本文件所揭露之演算法進行雲端資源之需求曲線的預測。 In step S430, the historical data is used to predict the cloud resource requirements of the first task. As disclosed in this document, the demand curve of cloud resources can be predicted through the algorithm disclosed in this document.
在步驟S440中,根據所預測之雲端資源需求,進行複數個雲環境之雲端資源配置。如本文件所揭露,在預測雲端資源需求之後,可分別將對應複數個雲環境的需求傳送出去,以達於複數個雲環境中配置對應之雲端資源。 In step S440, the cloud resource configuration of a plurality of cloud environments is performed according to the predicted cloud resource requirements. As disclosed in this document, after forecasting cloud resource requirements, the requirements corresponding to multiple cloud environments can be transmitted separately to achieve the corresponding cloud resources allocation in the multiple cloud environments.
第5圖繪示根據本揭示文件一些實施例中一種雲端資源管理裝置500的功能方塊示意圖。雲端資源管理裝置500包含辨識指令模組510、雲端資源需求項目取得模組520、資源組合計算與選定模組530、機器學習 模型540以及請求發送模組550。 FIG. 5 is a functional block diagram of a cloud resource management device 500 according to some embodiments of the present disclosure. The cloud resource management device 500 includes an identification instruction module 510, a cloud resource demand item acquisition module 520, a resource combination calculation and selection module 530, and machine learning Model 540 and request sending module 550.
如第5圖所示,辨識指令模組510用以接收一指令,或接收自動產生、因應事件產生、或週期性產生之一指令。關於接收或產生指令的說明如第2圖的步驟S210,於此不予重述。 As shown in FIG. 5, the identification instruction module 510 is used to receive an instruction, or an instruction that is automatically generated, generated in response to an event, or generated periodically. The description of receiving or generating an instruction is the same as step S210 in FIG. 2, and will not be repeated here.
雲端資源需求項目取得模組520用以擷取雲端資源需求項目。在一實施例中,雲端資源需求項目取得模組520根據辨識指令模組510辨識出的指令中之參數值,而獲得包含有雲端資源需求項目的所需要資源量和/或運算能力的資訊。在另一實施例中,雲端資源需求項目取得模組520向複數個雲環境發出請求以獲取關聯於指令的雲端資源需求之複數個雲端資源需求項目。 The cloud resource demand item acquisition module 520 is used to retrieve cloud resource demand items. In one embodiment, the cloud resource demand item acquisition module 520 obtains information including the required resource amount and / or computing power of the cloud resource demand item according to the parameter values in the instruction identified by the identification instruction module 510. In another embodiment, the cloud resource demand item acquisition module 520 sends a request to a plurality of cloud environments to obtain a plurality of cloud resource demand items associated with the commanded cloud resource demand.
資源組合計算與選定模組530用以根據指令所針對雲端資源需求項目的參數值計算出複數個資源組合。關於計算出複數個資源組合的說明如第2圖的步驟S230~S240,於此不予重述。 The resource combination calculation and selection module 530 is configured to calculate a plurality of resource combinations according to the parameter values of the cloud resource demand items targeted by the instruction. The description about calculating a plurality of resource combinations is shown in steps S230 to S240 in FIG. 2, and will not be repeated here.
機器學習模型540用以計算新的資源組合。在一些實施例中,機器學習模型540可針對過往的不同資源組合來執行數值演算法、機器學習演算法或其他演算法,以訓練出不同的模型,使得訓練出來的模型可用以計算新的資源組合。 The machine learning model 540 is used to calculate a new resource combination. In some embodiments, the machine learning model 540 may perform numerical algorithms, machine learning algorithms, or other algorithms for different combinations of resources in the past to train different models, so that the trained models can be used to calculate new resources. combination.
請求發送模組550用以針對機器學習模型540計算出的資源組合,向關聯於資源組合的一或多個雲環境送出請求。向雲環境發送請求的說明如第2圖的 步驟S250,於此不予重述。 The request sending module 550 is configured to send a request to one or more cloud environments associated with the resource combination for the resource combination calculated by the machine learning model 540. The instructions for sending a request to the cloud environment are as shown in Figure 2. Step S250 is not repeated here.
在一些實施例中,雲端資源管理方法200亦可實作為包含多個程式碼的電腦程式,並儲存於非暫態電腦可讀取記錄媒體。程式碼被載入至如第1圖的處理器120後,處理器120執行程式碼並操作如第2圖之步驟S210~S250。舉例來說,處理器120讀取指令以擷取雲端資源需求,經由網路介面110自複數個雲環境中獲得關聯於指令之複數個雲端資源需求項目。處理器120根據指令針對雲端資源需求項目之參數值以計算複數個資源組合。處理器120自該些資源組合中選定一資源組合,其中被選定的資源組合包含第一雲端資源以及第二雲端資源。經由網路介面110向第一雲環境送出第一請求,以於雲環境中配置第一雲端資源,以及經由網路介面110向第二雲環境送出第二請求,以於雲環境中配置第二雲端資源。 In some embodiments, the cloud resource management method 200 may also be implemented as a computer program including a plurality of codes and stored in a non-transitory computer-readable recording medium. After the code is loaded into the processor 120 as shown in FIG. 1, the processor 120 executes the code and operates as steps S210 to S250 in FIG. 2. For example, the processor 120 reads instructions to retrieve cloud resource requirements, and obtains a plurality of cloud resource requirement items associated with the instructions from a plurality of cloud environments via the network interface 110. The processor 120 calculates a plurality of resource combinations according to the parameter values of the cloud resource demand items according to the instructions. The processor 120 selects a resource combination from the resource combinations, and the selected resource combination includes a first cloud resource and a second cloud resource. Send a first request to the first cloud environment via the network interface 110 to configure the first cloud resource in the cloud environment, and send a second request to the second cloud environment via the network interface 110 to configure the second in the cloud environment Cloud resources.
在一些實施例中,雲端資源管理方法300亦可實作為包含多個程式碼的電腦程式,並儲存於非暫態電腦可讀取記錄媒體。程式碼被載入至如第1圖的處理器120後,處理器120執行程式碼並操作如第3圖之步驟S305~S335。 In some embodiments, the cloud resource management method 300 may also be implemented as a computer program including a plurality of codes, and stored in a non-transitory computer-readable recording medium. After the code is loaded into the processor 120 as shown in FIG. 1, the processor 120 executes the code and operates as steps S305 to S335 in FIG. 3.
在一些實施例中,雲端資源管理方法400亦可實作為包含多個程式碼的電腦程式,並儲存於非暫態電腦可讀取記錄媒體。程式碼被載入至如第1圖的處理器120後,處理器120執行程式碼並操作如第4圖之步驟 S410~S440。 In some embodiments, the cloud resource management method 400 may also be implemented as a computer program including a plurality of codes and stored in a non-transitory computer-readable recording medium. After the code is loaded into the processor 120 as shown in FIG. 1, the processor 120 executes the code and operates as shown in FIG. 4. S410 ~ S440.
在一些實施例中,非暫態電腦可讀取記錄媒體可為唯讀記憶體、快閃記憶體、軟碟、硬碟、光碟、隨身碟、磁帶、可由網路存取之資料庫或熟悉此技藝者可輕易思及具有相同功能之非暫態電腦可讀取記錄媒體。 In some embodiments, the non-transitory computer-readable recording medium may be a read-only memory, a flash memory, a floppy disk, a hard disk, an optical disk, a flash drive, a magnetic tape, a database accessible by a network, or a familiar This artist can easily think of non-transitory computer-readable recording media with the same functions.
綜上所述,本案提供一種雲端資源管理系統以及雲端資源管理方法,經由網路介面110向關聯於被選定的資源組合之雲環境發送請求,以於關聯的雲環境配置雲端資源,如此一來,即可實現雲端資源的動態配置。除此之外,本案可以根據當下的資源組合的運作狀況,例如業務需求、線上顧客數目、系統效能等因素,即時地判定狀況並動態地調整資源組合,以產生最適合雲端服務資源。 In summary, this case provides a cloud resource management system and a cloud resource management method. The network interface 110 sends a request to the cloud environment associated with the selected resource combination to configure cloud resources in the associated cloud environment. , You can achieve dynamic configuration of cloud resources. In addition, in this case, according to the current operating conditions of the resource combination, such as business needs, the number of online customers, system performance, and other factors, the status can be determined in real time and the resource combination can be dynamically adjusted to generate the most suitable cloud service resources.
上文概述若干實施例之特徵,使得熟習此項技術者可更好地理解本創作之態樣。熟習此項技術者應瞭解,可輕易使用本創作作為設計或修改其他製程及結構的基礎,以便實施本文所介紹之實施例的相同目的及/或實現相同優勢。熟習此項技術者亦應認識到,此類等效結構並未脫離本創作之精神及範疇,且可在不脫離本創作之精神及範疇的情況下產生本文的各種變化、替代及更改。 The features of several embodiments are summarized above, so that those skilled in the art can better understand the aspect of this creation. Those skilled in the art should understand that this creation can be easily used as a basis for designing or modifying other processes and structures in order to implement the same purpose and / or achieve the same advantages of the embodiments described herein. Those familiar with this technology should also realize that such equivalent structures do not depart from the spirit and scope of this creation, and can produce various changes, substitutions and alterations without departing from the spirit and scope of this creation.
Claims (8)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW108206991U TWM583564U (en) | 2019-05-31 | 2019-05-31 | Cloud resource management system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW108206991U TWM583564U (en) | 2019-05-31 | 2019-05-31 | Cloud resource management system |
Publications (1)
Publication Number | Publication Date |
---|---|
TWM583564U true TWM583564U (en) | 2019-09-11 |
Family
ID=68620432
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
TW108206991U TWM583564U (en) | 2019-05-31 | 2019-05-31 | Cloud resource management system |
Country Status (1)
Country | Link |
---|---|
TW (1) | TWM583564U (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110913016A (en) * | 2019-12-13 | 2020-03-24 | 杭州城市大数据运营有限公司 | Resource management system for cloud resource platform and storage medium |
TWI723410B (en) * | 2019-05-31 | 2021-04-01 | 伊雲谷數位科技股份有限公司 | Cloud resource management system, cloud resource management method, and non-transitory computer-readable storage medium |
-
2019
- 2019-05-31 TW TW108206991U patent/TWM583564U/en unknown
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
TWI723410B (en) * | 2019-05-31 | 2021-04-01 | 伊雲谷數位科技股份有限公司 | Cloud resource management system, cloud resource management method, and non-transitory computer-readable storage medium |
US11003503B2 (en) | 2019-05-31 | 2021-05-11 | Ecloudvalley Digital Technology Co., Ltd. | Cloud resource management system, cloud resource management method, and non-transitory computer-readable storage medium |
CN110913016A (en) * | 2019-12-13 | 2020-03-24 | 杭州城市大数据运营有限公司 | Resource management system for cloud resource platform and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
TWI723410B (en) | Cloud resource management system, cloud resource management method, and non-transitory computer-readable storage medium | |
US11656911B2 (en) | Systems, methods, and apparatuses for implementing a scheduler with preemptive termination of existing workloads to free resources for high priority items | |
CN113900767B (en) | Monitoring clusters and container-as-a-service controller implementing auto-scaling policies | |
US10514951B2 (en) | Systems, methods, and apparatuses for implementing a stateless, deterministic scheduler and work discovery system with interruption recovery | |
US11294726B2 (en) | Systems, methods, and apparatuses for implementing a scalable scheduler with heterogeneous resource allocation of large competing workloads types using QoS | |
US9571347B2 (en) | Reactive auto-scaling of capacity | |
US9106589B2 (en) | Predicting long-term computing resource usage | |
US10430218B2 (en) | Management of demand for virtual computing resources | |
US20150286507A1 (en) | Method, node and computer program for enabling automatic adaptation of resource units | |
US20160380908A1 (en) | Resource Prediction for Cloud Computing | |
US20140282520A1 (en) | Provisioning virtual machines on a physical infrastructure | |
US11726836B2 (en) | Predicting expansion failures and defragmenting cluster resources | |
US11064041B2 (en) | Apparatus for providing cloud service using cloud service brokerage based on multiple clouds and method thereof | |
TW201820165A (en) | Server and cloud computing resource optimization method thereof for cloud big data computing architecture | |
JP2015530676A (en) | Real-time optimization of computing infrastructure in virtual environment | |
CN110532086B (en) | Resource multiplexing method, device, system and storage medium | |
US11099893B1 (en) | Systems and methods for orchestrating seamless, distributed, and stateful high performance computing | |
US20180367631A1 (en) | Apparatus for providing cloud brokerage service based on multiple clouds and method thereof | |
US20230145025A1 (en) | Modeling cloud inefficiencies using domain-specific templates | |
TWM583564U (en) | Cloud resource management system | |
CN112000460A (en) | Service capacity expansion method based on improved Bayesian algorithm and related equipment | |
Naskos et al. | Cost-aware horizontal scaling of NoSQL databases using probabilistic model checking | |
CN112306666B (en) | Cloud resource management system and method, and non-transitory computer readable recording medium | |
Jayalakshmi | Predictive scaling for elastic compute resources on public cloud utilizing deep learning based long short-term memory |