Deployment and migration of business services in the cloud

Deployment and migration of business services in the cloud Autor: Bartosz Ignaczewski Director: Jose Luis Vazquez-Poletti Erasmus programme 2016-2017 1 Table of contents 1 Introduction .…………………………………………………………………… 5 1.1 Issue of the work ……..………………………..……………………….. 5 1.2 Scope and objective of the work ………..………………………………. 6 1.3 Structure of the work ……………………………………………………. 7 2 Deployment and migration of business services in the cloud .………………. 8 2.1 Cloud computing overview ………………………………………………8 2.1.1 Cloud computing definition ………………………………………. 8 2.1.2 Why do we need cloud computing? ………………………………. 9 2.1.3 Cloud computing characteristics …………………………………..11 2.1.4 Cloud computing deployment models …………………………….14 2.1.4.1 Public cloud …………………………………………….. 14 2.1.4.2 Private cloud ……………………………………………. 15 2.1.4.3 Community cloud ………………………………………. 15 2.1.4.4 Hybrid cloud ……………………………………………. 15 2.1.5 Cloud computing service models ………………………………… 16 2.1.5.1 Infrastructure as a Service ……………………………….17 2.1.5.2 Platform as a Service …………………………………….18 2.1.5.3 Software as a Service …………………………………….18 2.2 Overview of existing solutions ……….……………….…………………19 2.2.1 Deployment ………………………………………………………..19 2.2.1.1 Infrastructure as a Service ……………………………… 20 2.2.1.2 Platform as a Service …………………………………….21 2.2.2 Migration …………………………………………………………..22 3 Project CloudM ………………………………………………..………………..24 3.1 Project overview …………………………………………………………24 3.1.1 End users panel …………………………………………………….24 3.1.2 Administrator panel ………………………………………………..25 3.1.3 Sample use cases …………………………………………………..28 3.2 Tools, solutions and technologies used in the project ………………….. 29 3.2.1 Overview …………………………………………………………. 29 2 3.2.1.1 Ruby on Rails framework ……………………………… 29 3.2.1.2 Communication with Amazon Web Services ………….. 32 3.2.2 End users panel …………………………………………………... 32 3.2.3 Administrator panel ……………………………………………… 33 3.2.3.1 Amazon Relational Database Service section …………. 33 3.2.3.2 AWS Elastic Beanstalk section ………………………… 34 4 Project CloudM from the point of view of a user .……….……………..…… 37 4.1 End users panel …………………………………………………………….. 37 4.2 Administrator panel ………………………………………………………… 39 5 Conclusion……………………………………………………………………… 47 List of figures . ……………………………….….…………………………….. 49 List of code snippets ……………………………………………………………50 Bibliography ……………………………………………………………………51 3 4 Chapter 1 Introduction Cloud solutions in business are nowadays getting more and more popular. Many companies decide to deploy their applications in the cloud or migrate them from their non cloud based solutions. It allows them to focus on functionalities and turn over the work connected with software distributed architecture and scalability issues to cloud providers. 1.1 Issue of the work One of the most popular cloud providers are Amazon [1], Google [2] and Microsoft [3]. They allow clients to set up an account from which they can manage their infrastructures. Of course there are many ways to do it: for instance via their web-based online management tools, CLIs (Command Line Interfaces), APIs (Application Programming Interfaces). On top of them the companies and different communities built many SDKs (Software Development Kits) and libraries to speed up the process of administrating clients’ resources. As an example of those could be SDKs built by Amazon for the most popular languages and systems for the time being of writing this work [4]. However, the communities always build libraries which goals are to automate processes as far as they possibly can. So that, solely looking at the Ruby on Rails framework [5] to narrow the outcome, a couple gems (Ruby language [6] libraries) can be listed [7, 8, 9] to manage the AWS (Amazon Web Services) infrastructure, along with an application deployment, via special config files and CLIs. However, the competitive analysis has shown that there are no well-known possibilities of managing an application infrastructure together with a deployment and migration of business services via the application own administration panel. A user who wants to communicate with an AWS has to either have an terminal access or control everything through the Amazon website. Yet, such a solution is not well-centralized and 5 if a user wants to use multiple cloud providers, for instance because of better pricing that it has at some other provider for one of the modules he needs to use, he has to go to every provider’s website and make changes manually. Apart from being time consuming, which is costly because of both a time a worker has to spend to configure everything and passing time of using the unwanted cloud infrastructure, it is also error prone as such tasks are not automated, but reliable on a human. The solution to these issues is to create software that allows an administrator of an application to deploy and migrate business services for multiple cloud providers from only one panel in a simplified way. That brings another advantage: because of using an user-friendly UI (User Interface) and high level of automation instead of command line interface step-by-step approach, the actions can be controlled with a lower level of knowledge and training how specific building blocks are connected and working with each other. It directly translates to lowering labor costs that comes together with the ones mentioned earlier. This work presents a possible solution to the indicated problems. It is a web application written in Ruby on Rails framework with an administration panel that allows to deploy and migrate the system itself in the Amazon Web Services domain (along with creating a proper infrastructure). It is a prototype that in the future can be expanded to support more cloud providers and where more functionalities of a cloud base management can be added. 1.2 Scope and objective of the work The scope of this work is a field of software engineering, which is the design and implementation of a web application with the use of available programming framework and libraries. The objective of this work is to design and implement a web application that allows to deploy and migrate its parts to the cloud to achieve high availability and scalability of the system using the Ruby on Rails framework. 6 1.3 Structure of the work The first section of this work provides a brief introduction to the subject of the cloud along with a brief presentation of the needs of its existence. It includes the scope and objective of the work. The second chapter includes a general overview of the cloud along with application deployment and migration solutions. It presents selected solutions and points out their main advantages and disadvantages. The third chapter presents an accurate description of the created web application. It consists of a delineation of the functional requirements of the system and what tools and technologies were used to build it. The fourth chapter contains an extensive user documentation. It is a guide, introducing user, how it can use the built website. The fifth chapter is a summary of all the work.! ! ! ! ! ! ! 7 Chapter 2 Deployment and migration of business services in the cloud The Software Development Process Life Cycle consists of many steps at the end which we have phases of deployment and maintanence, which includes migration [10]. As it is extremely crucial to focus on building great, reusable software it is also essential to remember the importance of these two last aspects, on which the cloud computing focuses. 2.1 Cloud computing overview 2.1.1 Cloud computing definition ! There are many definitions of what exactly cloud computing is. The National Institute of Standards and Technology definition [11] is as follows: „Cloud computing is a model for enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction. This cloud model is composed of five essential characteristics, three service models, and four deployment models.”. On the other hand, the Amazon definition [12] is described as: „Cloud computing is the on-demand delivery of compute power, database storage, applications, and other IT resources through a cloud services platform via the internet 8 with pay-as-you-go pricing.”.! Looking at above it can be said that cloud computing is the on-demand service that allows management of various IT resources via the Internet which has certain characteristics and models. As the characteristics and models are described in detail in the following sections, first a question „why do we need cloud computing?” should be answered. 2.1.2 Why do we need cloud computing? ! Cloud computing, as a solution, is a response to some problems that occur during the development process in the (especially) last years. Its advantages are the best seen in large project, which require for instance high availability, performance and scalability. As a software is being created and deployed for the first time it is usually done on a simple server that hosts the application (actually, there shall be at least two to avoid a single point of failure). The cost of usage does not change and monitoring can be done easily. As the project grows, gains more and more users and manages lots of data (getting to sizes of big data) the administrator of the server starts to spend an increasing amount of time on maintenance, monitoring. Assuming it was not done at the beginning, at some point there is a need to scale as the high performance and availability start to be endangered. There are two ways an administrator can go to scale: 1. Horizontally - add more servers to the existing ones and connect them, so that processing the workload is now distributed among all of them 2. Vertically - add more resources within the existing server, which basically means increasing the RAM amount or CPU capabilities; improving hardware specification in general The main next things that are going to happen are (at least): 9 However, businesses have to take care of a reliable and fast connections between data centers security policies compatibility. In conclusion, different types of cloud allow businesses to choose which of them suits the best their needs. They provide incrementally more and more scalability, reliability, cost-efficiency and availability in trade of privacy and data protection in comparison to a „classic” own data center approach. Figure 4 presents their popularity. Figure 4: Cloud computing trends in 2016 [18] Apart from virtualization, from which the discussion about privacy and security was started, cloud computing also relies heavily on service-oriented architecture (SOA). It means that cloud providers offer services which clients can utilize. However, there are different models, which are described in the following section. 2.1.5 Cloud computing service models According to the National Institute of Standards and Technology definition, there are three different cloud computing service models: Infrastructure as a Service (Iaas), Platform as a Service (Paas) and Software as a Service (Saas). They are reffered 16 as the „cloud stack”, because they are increasingly providing more comprehensive levels of abstraction. Figure 5 is their visual representation. Figure 5: Visual representation of the cloud stack [19] 2.1.5.1 Infrastructure as a Service Infrastructure as a Service, abbreviated as IaaS, is the least abstract layer of the cloud computing stack. It is mostly used by administrators as it conceptually allows to manage IT resources (infrastructure: for instance servers, storage, network components, operating systems and load balancers) with a great level of control. It directly uses virtualization to emulate proper environments which have to be handled manually. This is a model that was referenced to the most through previous chapters and sections. Practically, it is designed to be used by administrators responsible for building the whole system infrastructure from the start, perform software updates, etc. 17 2.1.5.2 Platform as a Service Platform as a Service (Paas) provides an incremented level of abstraction in comparison to the IaaS model and it is designed to help developers utilize cloud by serving „out-of-the-box” infrastructure, operating system and various technologycentered execution environments so that web applications can be deployed in minutes without (almost) any additional work. It also takes care of other phases of the software development process. For instance, a developer can set up a continuous integration service for testing purposes. Theoretically in this model a loss of control in comparison to the IaaS type can be seen. However, many cloud providers allow managing the infrastructure despite using the PaaS model, as it is a „higher” layer on the stack. The major issue with this model is a possibility of a vendor lock-in. It is one of the possible disadvantages of the cloud that was not mentioned before. Generally, when a client uses services from one cloud provider it may be very difficult to change to services offered by another one. This might be caused by for example different technologies usage or various incompatible implementations of core functionalities. 2.1.5.3 Software as a Service Software as a Service (SaaS) is a model designed to be utilized by end users, that do not have any technical skills. Google Apps, iCloud could be shown as examples. The real control is very narrow and users are usually presented a very limited set of settings, but do not have to take care of software updates and similar tasks (for instance scaling) at all. Usually, SaaS offers subscription-based services in comparison to a „on-demand” „pay-per-use” IaaS model. There are APIs that allow integration between different systems. Vendor lock-in is here very easily seen. Apart from the three models mentioned in the National Institute of Standards and Technology definition there are also others, for instance Network as a Service, Identity as a Service, Mobile Backend as a Service, but describing them is out of scope of this work. 18 This work focuses entirely on a public cloud model due to little or no expenses required to do so as well as Amazon as a provider because of the highest popularity and great community support (figure 6). Figure 6: Public Cloud Adoption 2016 vs 2015 [18] 2.2 Overview of existing solutions Deployment and migration in the cloud can be done in multiple ways. Not only Amazon provides tools to do so (like a web interface, command line interface, SDKs), but there are also many libraries and other solutions managed by a vast open-source community and enterprises. 2.2.1 Deployment When focusing on a deployment it is very important to answer which service model we are going to use. As it may differ depending which one we choose, the following sections cover both Infrastructure as a Service and Platform as a Service. 19 2.2.1.1 Infrastructure as a Service Amazon offers (almost) countless services to support its cloud computing solutions. However, their basic resource that is being used in the infrastructure is Amazon Elastic Compute Cloud (EC2) [20] which presents itself as a virtual server. To allow quick deployment and usage it offers predefined templates of operating systems and environments that we can choose to be automatically installed on our instances. These templates are known as Amazon Machine Images (AMIs) [21]. They can be easily customized (and then copied to other EC2 instances) and shared with other people. For example, if we would like to have support for a very customized software stack, we can look for it and there is a great chance that either Amazon or the community has already published one. If there is none, we could prepare it once and then share it with others. It greatly improves the time we need to spend on different configurations. There are many solutions to support more automatic and easier deployment of EC2 instances. Amazon by themselves offers Amazon OpsWorks [22] which allows to use Chef [23] in order to build infrastructure by specifing everything with the code. An alternative to this is using Puppet [24], which also grants a possibility to „code” a cloud infrastructure and then run scripts that build it. The mentioned tools could be used in more ways (like a continuous integration with Chef), but it is outsite of scope of this work to present them all. An example of a Puppet code (code snippet 1): Code snippet 1: Puppet code example to deploy an EC2 instance [25] 20 ec2_instance { 'name-of-instance': ensure => present, region => 'us-east-1', availability_zone => 'us-east-1a', image_id => 'ami-123456', instance_type => 't1.micro', monitoring => true, key_name => 'name-of-existing-key', security_groups => ['name-of-security-group'], user_data => template('module/file-path.sh.erb'), tags => { tag_name => 'value', }, } Focusing more on a Ruby on Rails framework which will be used as a main framework to build a website that will be deployed on the Amazon cloud computing platform, an equivalent solution to the presented ones could be a rubber gem [26]. EC2 also supports the more „classical” approach, so that if we have a project with a configured version control system like Git [27], we can access the server via a terminal, pull the changes and restart the server. Another way is to use special deployments gems that work the same as on a non-cloud based machines [28]. These solutions are great, but as stated earlier, Infrastructure as a Service model requires a lot of work and knowledge to set up and it is designed mostly to be used by administrators. Code changes that are required to be made by Chef or Puppet are enormous and take time, the „classical” approach does not build an infrastructure (which has to be done manually on the AWS website for instance) and in the end it does not automate the process of deployment. Not mentioning not having a proper user interface so that could be done inside a web application. However, Infrastructure as a Service is not the only model that is used in cloud computing. 2.2.1.2 Platform as a Service As this work focuses on building a web application, Amazon recommend using its Platform as a Service service AWS Elastic Beanstalk [29]. Its advantages are enormous. As it is designed to be mostly used by developers, it automates building the infrastructure to the level that only application deployment is necessary and the rest is handled by Amazon. It automatically adds load balancers and other components required to properly run the project. Moreover, a user still has full control of every resource it is used, as AWS Elastic Beanstalk is built on top of EC2 instances and other infrastructure blocks known from the Infrastructure as a Service model, with its advantages. Time needed to deploy a fully working application is greatly reduced. It can be done in multiple ways: by uploading a .zip file, using command line interface or a special one specifically created to be used with Elastic Beanstalk [30]. There are even Ruby libraries to make things easier framework-wise [31, 32]. 21 However, this solutions also do not have a user interface which could simplify things even more. A user has to be a skilled technical person to configure and run specified commands nonetheless. In addition, all presented ways of deployment are only focusing on AWS. The application built in this work concentrates on being expanded to have a possibility to support many cloud providers. That being said, as it can use tools provided by specific providers in its underlying implementation, it cannot be fully relying on them. 2.2.2 Migration Migration, as deployment, can be done in multiple ways. Practically, it comes down to a database creation, applying a schema that is present on a one to be replicated and copying data. As creation could be done during the deployment phase, this section focuses mostly on data replication. Data can be migrated either with Amazon specific interfaces or using general solutions. The latter could be using rsync [33] (which Amazon recommends) that allows to transfer files or, for instance, dumping the whole database to run scripts on the other one. Although it is possible, when there is a huge data center and databases have to be available all the time or for the sake of convenience we would like to use a direct solution, Amazon provides many services do to so. Depending on what exacly we need, we could either transfer petabytes of data, connect directly to a reginal datacenter, use a Schema Conversion Tool (if we migrate from one type of database to another) or run some commands from the command line interface [34, 35]. The topic of data migration is extremely complicated and the possibilities are so many, mostly not free, that describing all of them is well outside of the scope of this work. However, the important disadvantage of all of them is that presented solutions do not provide an easy way of data migration without a specific vendor lock-in (in this case, Amazon). There is not a well-known service that could integrate with Amazon to migrate for free the existing data within the web application’s user interface. 22 Because of the specified lacks of functionalities that were spotted in the existing solutions, mainly either vendor lock-in, not a free service or high complexity without a proper user interface inside an own web application to manage resources, this work presents a prototype project that integrates with Amazon Web Services and that allows easy deployment and migration. Ideally, it is designed to support multiple cloud providers and many advanced options, which could be added in the future. 23 Chapter 3 Project CloudM Project CloudM was built with an idea that an application administrator, even with almost no knowledge of cloud computing or computer science in general, could use it after a short introduction of the action flow in the settings panel. It is a prototype that integrates with Amazon Web Services and allows extremely simplified deployment and migration of business services and data. 3.1 Project overview The project consists of two main panels: end users and administrator panel. The first one is designed to be used only by „real” application users, whereas an administrator can change application settings, connect with the cloud and create an infrastructure to support it. This section presents main functionalities and requirements used in the prototype in a high-level description. The more technical approach will be shown in the next part of this work as well as a user guide with corresponding screenshots. 3.1.1 End users panel The end users panel is an example of an application that could be used in a real world. It is a microblogging tool with a possibility to CRUD (Create, Read, Update, Delete) posts that consist of title and body. When a post is being created or updated, a user has to provide these two fields as the validator checks if they are not blank, otherwise displaying a message that there are errors in the form and the post cannot be saved. The index page presents a list of posts that are currently in a database. To ease the navigation there is a pagination with a default 10 posts per page setting set. 24 Before any post deletion there is an additional pop up window asking if a user is sure about this decision. As this panel is only an example where the main goal is to provide data that could be migrated to a cloud database later on, there is no need to implement more advanced functionalities like filtering posts or linking to other models. However, it is important to mention that the application is built in a way that can be easily developed in any direction in the future. The most important part of this prototype is directly connected to the cloud infrastructure management, and the administrator panel allows it. 3.1.2 Administrator panel Before using actions linked with changing any settings of the application a user has to log in to have privilleges to do it (posts CRUD actions do not require that). However, there is one method that can be triggered without a session: a superuser creation. The idea why it is done like this is very simple: when we create a new database in a cloud and connect to it, there is no data. To migrate data we need to go to the settings page, which is inaccessible without being logged in. As we do not have any user account created in the new database, we can do it in multiple ways: within the project interactive shell, or the database console which can be directly accessed via the SSH, but the idea of this prototype is to be as simple as possible and do things as quickly as possible. Because of that, there is a button that allows a superuser creation so that it is managable to log in within seconds after connecting to a database. After a user logs in and goes to the settings page there are new options available. There are two sections that provide actions connected with the application deployment and migration: the Amazon RDS (Relational Database Service) [36] section and the AWS Elastic Beanstalk section. The first one is responsible for actions linked with database creation, deletion and migration, the latter with application deployment and deletion. As it has been already described in the previous chapters, AWS Elastic Beanstalk is a PaaS solution recommended to developers that allows cloud infrastructure management on a higher level than creating each component separately. 25 3.2.1.2 Communication with Amazon Web Services There are many ways to manage Amazon Web Services components. It can be done for instance via the command line interface, directly through their website, mobile application or the REST API which can be used in the Ruby on Rails framework project by adding the AWS SDK gem [44]. This basically gives access to all possible services and the AWS SDK documentation profoundly describes them. In the CloudM prototype I decided to use the REST API, as it is the most common way to integrate with other systems and it is officially supported by Amazon. However, to use the SDK we need to set up a new account on Amazon, create special keys that will be used for authentication and pass them in the requests. As in the requirements for the prototype there is no action connected with giving the API keys (because storing passwords is a whole different topic to address and it is not easy to do), Amazon’s documentation [44] actually mentions that the best way to do it is to create a secret file that contains them and is ignored by our source control system (Git [45] in this case). As the administrator panel uses only one account and there are no more accounts, it is a perfect way to add the API keys to be authenticated by Amazon. Using Git as a source control has many advantages which describing is out of scope of this work. Despite that, there is one special favor linked to it which will be presented in one of the next sections. As it is clear what is the main framework and a way of communication with Amazon Web Services in the CloudM prototype, there is a need to focus on more specific tools and solutions that are implemented. The description is divided into two parts: one for the end users panel and the other one for the administrator panel, which itself has two more sections. 3.2.2 End users panel End users panel presents CRUD actions along with corresponding views. It was built with a modification of a default scaffold template. First, Bootstrap [46] was used to 32 gain interactive, reusable components to create the user interface. Then, some other gems (Ruby libraries) were put into the prototype: -gem 'bootstrap-sass' [47] : to add Bootstrap -gem 'haml-rails' [48] : to support HAML markup language [49] instead of the default one -gem 'simple_form' [50]: a Bootstrap integration with forms in views -gem 'bootstrap-generators' [51]: a Bootstrap integration with scaffolds -gem 'kaminari' [52]: allows adding pagination in views for models -gem 'bootstrap-kaminari-views' [53]: a Bootstrap integration with Kaminari On top of that, the whole project uses gem 'growlyflash' [54] which is responsible for notification (flash) messages. 3.2.3 Administrator panel Apart from Bootstrap, Simple Form, Growlyflash and HAML that are used in the project globally, administrator panel also implements mentioned gem 'aws-sdkrails' (for AWS REST API connections), gem 'devise' (for authentication) [55] and a gem 'devise-bootstrap-views' (for Bootstrap integration with Devise) [56]. As the administrator panel consists of two parts, each of them has its own solutions and additional gems that were used to build it. 3.2.3.1 Amazon Relational Database Service section This section contains information and actions linked to a database connection, data migration and the cloud infrastructure management. Database creation is based on sending specific parameters to Amazon Web Services: the engine is postgres [57] and, what is important when using a free tier, the instance class is 'db.t2.micro'. Other key - value pairs contain information about istance identifier, name, user name and password. 33 After the database was created there is a need to run schema migrations. This is the time when the Ruby on Rails framework is in favor: to do it, all what has to be done is invoking a 'db:migrate' command. It looks similar when it goes to preparing and running data migrations. Ruby on Rails has special files and tasks to support it. As for the data preparation a gem 'seed_dump' [58] was used to dump currently held data to a file, whereas invoking migrations is as simple as running a 'db:seed' task (Code snippet 5). Code snippet 5: Invoke data migration method Other methods, for instance connecting to a database (local and Amazon RDS), its deletion and getting its information are implementations that use either AWS REST API or Ruby on Rails methods. 3.2.3.2 AWS Elastic Beanstalk section When creating and deploying the project to the cloud it is important to upload a current version of the application. This is done with a help of two solutions (apart from the REST API): Git and AWS CodeCommit [59], which allows to store the Git repository. In the first part of the project deployment (application version creation), we need to specify how we are going to upload the project. It could be either a *.zip file (which can be also generated with Git [$ git archive --format=zip HEAD]), or an AWS CodeCommit repository. Since generating a *.zip file connects with uploading it to some server what can take a lot of time, I decided to use the AWS CodeComit service. Creating an application version in this case is straightforward and it is based on a created repository (Code snippet 6). 34 Code snippet 6: Create application version method As the application and its version are created, now it is time to create the environment which will have an endpoint that will allow to access it. It is extremely impotant to remember that in the production environment we are using environment variables (Code snippet 2) that are referencing information about the database connection details. Without them (and an existing database) some scripts will not work (for instance those responsible for schema migrations) and the whole deployment will fail. To generate them, we need to create special options (Code snippet 7). Code snippet 7: Generate database environment variables method 35 There are also other parameters that are being sent to Amazon when making a request, for example the solution stack name which tells what type of system and preinstalled software we would like to have on a server (as AWS Elastic Beanstalk is a PaaS solution) or the instance type (Code snippet 8). Notice how we use the generate_db_opts method that is presented earlier (Code snippet 7). Code snippet 8: Environment creation parameters 36 Chapter 4 Project CloudM from the point of view of a user As the application is designed to be used by two different types of users, this chapter is divided into two sections: one that is designed to be read by end users and the other one is meant to be for administrators. Ideally, the prototype would already run in the cloud (with a provider different than Amazon), but for the purpose of this work it is running locally. 4.1 End users panel When a user enters the website, the main Microblog application is presented (Figure 8). Figure 8: The Microblog application main page 37 2 3 4 1 5 It consists of posts listed by descending Created at attribute (1). There are also buttons and links to Create (2), Update and Delete (3) actions. On the bottom left hand corner there are pagination links (4). By default, there are ten posts listed per page. As the user wants to create a new posts, it clicks a New Post button (2) and it is redirected to another page (Figure 9). Figure 9: The New post page To create a new post it is required to fill in the two text fields: Title and Body. If there is no content in any of them, there is an error as there is a validation on a Post model (Figure 10). Figure 10: The New post page validation errors: can’t be blank for the Title and Body fields After a user enters values into both of the text fields, it is redirected to a main page with a notification (flash message) in the top right hand corner (1) about the successful action (Figure 11). Notice the new post listed as a first one (2). 38 Figure 11: New post was created The Edit action redirects to a Post form shown on Figure 9, but the difference is that the values of the current post are already filled in the text fields. The Destroy action performs a post deletion, but before that a user is asked if it is sure that it wants to remove it (Figure 12). Figure 12: The confirmation dialog for a post deletion 4.2 Administrator panel There are two additional links in the navigation bar present on Figure 8 (5): AppSettings and Sign in, which are designed to be used by an administrator. To manage the settings and configuration of the cloud infrastructure and the application itself, an administrator has to log in first. After clicking Sign in, it is 39 1 2 redirected to a form where it can do that (Figure 13). Additionally, it can check the Remember me checkbox (1) to hold a session. Figure 13: The administrator log in form Of course, if a user name and password do not match, there is a flash message informing about it. When an administrator is logged in, it has a full control over the settings. Assuming that the application was never deployed to the cloud, basic actions are available (Figure 14). Figure 14: The administrator’s AppSettings page 40 1 1 2 3 Apart from the Create database button (1), there are some information how the application has to be used. For instance, to deploy the application a database has to be created first (2) and to get a new status information the page has to be refreshed (3). After the Create database action was triggered, new options became available (Figure 15). Figure 15: The AppSettings page after clicking the Create database button First, there is a notification about the action taken (1). Then there is new information about the database: its status and an endpoint (as Amazon did not set it up yet, it is blank) (2). The administrator can see to which database the application is currently connected (3) and it can Prepare data migrations (4) that can be used later on to migrate current data to the AWS RDS. Notice that the Sign in link is now replaced by Sign out one (5): it is present since the administrator logged in. After a while the page can be refreshed and it can be seen new status, endpoint information (1), possibility to connect to the AWS database (2) and to delete it (3) (Figure 16). When the administrator connects to the newly created database, it cannot log in because there is no data on it. To create the administrator account it can go to AppSettings, which has an action to do it (Figure 17). If the administrator wants to see on AWS RDS website the information about the newly created database, it can do it (Figure 18). Notice the endpoint address (1), one current connection (2), the instance class (3) and the status (4). 41 1 2 3 4 5 detailed level. There are many advantages of this technology, especially in comparison to some existing solutions. The Ruby on Rails framework would not exist if it was not for the Ruby language, which allows extremely fast functionalities creation mainly due to the fact that it requires very little code (in comparison to Java for instance) to implement them. Thanks to the mentioned technologies the CloudM prototype was created and it consists of many operations that allow the management of the cloud infrastructure and data migrations. 48 List of figures [1] The infrastructure capabilities versus the actual demand [2] How virtualization works [3] Visual representation of cloud computing models [4] Cloud computing trends in 2016 [5] Visual representation of the cloud stack [6] Public Cloud Adoption 2016 vs 2015 [7] AWS Elastic Beanstalk logs from an environment creation [8] The Microblog application main page [9] The New post page [10] The New post page validation errors: can’t be blank for the Title and Body fields [11] New post was created [12] The confirmation dialog for a post deletion [13] The administrator log in form [14] The administrator’s AppSettings page [15] The AppSettings page after clicking the Create database button [16] Visible options after the database creation (with the available status) on the AppSettings page [17] The AppSettings Create a superuser account button and the information about it [18] The AWS RDS database information [19] The AppSettings page after connecting to the AWS database [20] The Create application action button (after the full database creation process) [21] The Create environment action button [22] The AWS Elastic Beanstalk section after full application deployment [23] The AWS Elastic Beanstalk console information [24] Fully deployed application is available at the endpoint URL [25] No Delete database action button when the application is fully deployed on the AppSettings page [26] The Terminate application button shows up only when the environment status is set to Terminated 49 List of code snippets [1] Puppet code example to deploy an EC2 instance [2] Ruby on Rails database connection configuration code [3] Ruby on Rails scaffold generation [4] Getting the current date and time in Java [5] Invoke data migration method [6] Create application version method [7] Generate database environment variables method [8] Environment creation parameters 50 Bibliography [1] Amazon AWS main website https://aws.amazon.com, 09.02.2017 [2] Google Cloud Platform main website https://cloud.google.com, 09.02.2017 [3] Microsoft Azure https://azure.microsoft.com, 09.02.2017 [4] Amazon tools for developing and managing AWS applications https:// aws.amazon.com/tools/?nc1=f_dr, 09.02.2017 [5] Ruby on Rails framework website http://rubyonrails.org, 09.02.2017 [6] Ruby language website https://www.ruby-lang.org, 09.02.2017 [7] Rubber gem github webpage https://github.com/rubber/rubber/wiki, 09.02.2017 [8] Puppetlabs gem github page https://github.com/puppetlabs/puppetlabs-aws, 09.02.2017 [9] Elastic Beanstalk gem github page https://github.com/alienfast/elastic-beanstalk, 09.02.2017 [10] Software Development Process https://en.wikipedia.org/wiki/ Software_development_process, 09.02.2017 [11] The National Institute of Standards and Technology Definition of Cloud Computing, http://nvlpubs.nist.gov/nistpubs/Legacy/SP/ nistspecialpublication800-145.pdf, 09.02.2017 [12] AWS Cloud Computing intruduction, https://aws.amazon.com/what-is-cloudcomputing/, 09.02.2017 [13] LinkedIn blocked in Russia news, https://www.theguardian.com/world/2016/nov/ 17/russia-blocks-access-to-linkedin-over-foreign-held-data, 10.02.2017 [14] The difference between virtualization and cloud computing, https:// blog.cloudpassage.com/2011/07/22/the-difference-between-virtualization-and-cloudcomputing/, 10.02.2017 [15] AWS security resources, https://aws.amazon.com/security/security-resources/, 10.02.2017 [16] Information about cloud computing, https://en.wikipedia.org/wiki/ Cloud_computing, 10.02.2017 [17] Information about cloud deployments models, http://www.slideshare.net/melsatar/ cloud-deployments-models, 10.02.2017 51 [18] Information about cloud computing statistics, http://www.rightscale.com/blog/ cloud-industry-insights/cloud-computing-trends-2016-state-cloud-survey, 10.02.2017 [19] Information about cloud stack, https://web.archive.org/web/20151121220939/ http://www.clouderpc.com/making-sense-of-the-cloud-saas-paas-and-iaasexplained#.WJ79vSTiTjA, 11.02.2017 [20] Amazon EC2 website, http://docs.aws.amazon.com/AWSEC2/latest/UserGuide/ concepts.html, 12.02.2017 [21] Amazon Machine Image website, http://docs.aws.amazon.com/AWSEC2/latest/ UserGuide/AMIs.html, 12.02.2017 [22] Amazon OpsWorks website, https://aws.amazon.com/opsworks/, 12.02.2017 [23] Chef website, https://www.chef.io, 12.02.2017 [24] Puppet website, https://puppet.com, 12.02.2017 [25] Puppet Rails gem website, https://github.com/puppetlabs/puppetlabs-aws, 12.02.2017 [26] Rubber gem website, https://github.com/rubber/rubber/wiki, 12.02.2017 [27] Git website, https://git-scm.com,12.02.2017 [28] How to deploy Ruby on Rails application to AWS website, https:// www.sitepoint.com/deploy-your-rails-app-to-aws/, 12.02.2017 [29] AWS Elastic Beanstalk documentation website, https://aws.amazon.com/ documentation/elastic-beanstalk/, 12.02.2017 [30] Deploying a Ruby on Rails application on AWS Elastic Beanstalk website, http://docs.aws.amazon.com/elasticbeanstalk/latest/dg/create_deploy_Ruby_rails.html, 12.02.2017 [31] Elastic Beanstalk Ruby gem website, https://github.com/alienfast/elastic-beanstalk, 12.02.2017 [32] Elastic Beanstalk deployer Ruby gem website, https://github.com/ ThoughtWorksStudios/eb_deployer, 12.02.2017 [33] rsync website, https://rsync.samba.org, 12.02.2017 [34] AWS Cloud Data Migration website, https://aws.amazon.com/cloud-datamigration/, 12.02.2017 [35] AWS Database Migration Service website, https://aws.amazon.com/dms/? nc1=h_ls, 12.02.2017 52 [36] Amazon RDS docs website, http://docs.aws.amazon.com/AmazonRDS/latest/ UserGuide/Welcome.html, 16.03.2017 [37] Ruby on Rails framework website, http://rubyonrails.org, 17.03.2017 [38] Ruby language website, https://www.ruby-lang.org/en/, 17.03.2017 [39] Basecamp website, https://basecamp.com, 17.03.2017 [40] Convention Over Configuration, https://en.wikipedia.org/wiki/ Convention_over_configuration, 17.03.2017 [41] JSON documentation, https://www.w3schools.com/js/js_json_intro.asp, 17.03.2017 [42] Don’t Repeat Yourself, https://en.wikipedia.org/wiki/Don%27t_repeat_yourself, 17.03.2017 [43] How to get the current date and time in Java language, http://www.mkyong.com/ java/java-how-to-get-current-date-time-date-and-calender/, 17.03.2017 [44] AWS SDK for Ruby docs, http://docs.aws.amazon.com/sdkforruby/api/index.html, 17.03.2017 [45] Git website, https://git-scm.com, 17.03.2017 [46] Bootstrap website, http://getbootstrap.com, 17.03.2017 [47] Bootstrap-saas gem website, https://github.com/twbs/bootstrap-sass, 17.03.2017 [48] Haml-rails gem website, https://github.com/indirect/haml-rails, 17.03.2017 [49] HAML markup language website, http://haml.info, 17.03.2017 [50] Simple_form gem website, https://github.com/plataformatec/simple_form, 17.03.2017 [51] Bootstrap-generators gem website, https://github.com/decioferreira/bootstrapgenerators, 17.03.2017 [52] Kaminari gem website, https://github.com/kaminari/kaminari, 17.03.2017 [53] Bootstrap-kaminari-views gem website, https://github.com/matenia/bootstrapkaminari-views, 17.03.2017 [54] Growlyflash gem website, https://github.com/estum/growlyflash, 17.03.2017 [55] Devise gem website, https://github.com/plataformatec/devise, 17.03.2017 [56] Devise-bootstrap-views gem website, https://github.com/hisea/devise-bootstrapviews, 17.03.2017 [57] Postgres database website, https://www.postgresql.org, 18.03.2017 [58] Seed_dump gem website, https://github.com/rroblak/seed_dump, 18.03.2017 53 [59] AWS CodeCommit website, http://docs.aws.amazon.com/codecommit/latest/ userguide/welcome.html, 18.03.2017 54