Cloud computing system has some functional and non-functional properties, but at the same time it has many interesting features. The relevant issues have to address the cloud providers to use and verify that the target system requires some of the performance levels in SaaS, PaaS, Iaas. Most of the cloud systems are multi-capable and varies performances within a day and holds congestion in the network. In this work, the authors stimulate on some of the attainment issues, which will maintain the benefit of multi-cloud users. In the era, period of multi-cloud has been long promoted to the standpoint of administrative efficiency. Some of the CSPs offer nice perks and motivated their services under one umbrella- a term which provides cloud applications from multiple cloud computing providers known as Multi-Cloud computing. In 2016, a report from BCN (Business Cloud News) says 57% enterprises had no multi-cloud approach, 35% do not have private cloud design, and 28% lack for a public cloud design at all. A source from Righscale-2018, state of the cloud says 82% utilizes the Multi-Cloud, 9% uses single public cloud, 4% uses single private cloud, and 5% had no plans were shown in Figure 1 (Rowan, 2017).

Figure 1. Rise of Multi-Cloud

The aim of Multi-Cloud (Senthil, 2012) is to merge multiple CSPs and has a different set of API to handle the various network storage architecture. The benefits of Multi-Cloud are redundancy, accuracy, scalability, fast response time, unique cloud-specific, public cloud cost benefits, utilize the existing investments, and disaster recovery. A Multi-Cloud strategy needs a new capability of service, it does not limit on one cloud provider's offerings. It makes a sense of reliability to meet a high - level business, when comes to Multi-Cloud interoperability is the key, makes a view to single management, and monitoring tool.

A Multi-Cloud approach (Rowan, 2017) provides organizations and developers to pick the component that will comprise the application with workloads on demands. There are some few technical barriers to meet the specific services that need to attain on Multi-Cloud. Cost management should be the priority from the start initially from moving workloads depending on the specific application (TechTarget, 2019). There are eight ways for an enterprise to follow in investing Multi-Cloud:

• Investigate all available options – They need to understand and evaluate their needs to the provider's best pricing.
• Map-out your strategy – Evaluate all of the available options and pick the most suitable for the workload which is needed.
• Fully analyze business requirements – Organizations deploy application components to business requirements that can reduce data hops across the cloud boundaries.
• Use insightful, reliable monitoring – Continuously monitoring infrastructure maintains that resources and money aren't being wasted in a Multi-Cloud environment.
• Ensure proper data classification and tiering – It is essential for Multi-Cloud cost management. Segregation of data and expensive data transfer across cloud platforms should be handled carefully.
• Don't be short-sighted – Make proper planning to secure buy-in business units.
• Exercise strong management – It gives an organization tremendous flexibility to deploy the mission-critical applications.
• Don't scrimp on security – Failing to understand CSPs are not responsible for stored data assets, add-on security for our application deployment is needed.

Most of the organizations find services from cloud providers are affordable when contrasted with the investment cost of physical servers and system infrastructures. In the wake of picking up involvement with single providers currently, they are moving towards the multi-cloud environment to receive the rewards of cost, execution, and unwavering quality. To have the option to utilize services over different cloud providers requires the utilization of control and the executive's stage that would enable organizations to send and deal with their applications over the best providers and best services that meet their destinations. They should be upgraded to perform well. For improvement to succeed, it must be gone before by profiling and execution assessment. This exploration will accomplish the procedure of performing various tasks condition in the multi-cloud foundation by having some powerful tools for estimating overall presentation and advancement of multi-cloud computing infrastructures. The goal of this article is to deploy a multi-cloud environment that mirrors the best in the class of Cloud service and evaluating the performance and makes the user involves in Multi-Cloud. In this work, the authors provide an analysis of various Multi- Cloud applications produced for services obliged users to check their analysis based on parameters as CPU utilization, storage use, and time utilized.

The cloud computing terminologies (Guerriero, Ciavotta, Gibilisco, & Ardagna, 2015) allow us to configure systems that are able to opt their performance with delegate cloud providers. To maintain the cloud infrastructure that would like to have with our needs with required Service Level Objectives (SLOs). The model-driven approach was presented to attain evaluation and cost proximity of a single cloud and Multi-clouds. Some works (Franceschelli, Ardagna, Ciavotta, & Di Nitto, 2013; Sefraoui, Aissaoui, & Eleuldj, 2012) discussed with opensource software for cloud implementation and make the users in selecting opensource software for their best solution requirement. Another work (Jain, Kumar, & Anamika, 2014) proposed Space4Cloud, a Dev-Ops provides an integrated environment for optimization and runtime capacity allocation for a cloud application. The variations of timestamp suggested with the varying number of users and Virtual Machines (VMs) allotted in it. Another work (Gupta, Jain, Samaka, Erbad, & Bhamare, 2016) proposed in finding hiring resources from CSP had been less expensive, when it was compared to buying physical servers and network equipment. They discussed with OpenADN, a Multi-Cloud management system, enterprises/organizations prompting to utilize resources on multiple cloud providers. This work addressed with two sets of parameters affecting CPU time and compared the effects of functions like hosting, polling, and sleep workload performance with varying number of users and number of hosts. Indira and Devi (2014) focussed on CPU scheduling, memory usage, I/O tasks, and resource sharing on Multi-Cloud architectures. VM utilizes the resources instead of a real system, and this VM utilizes the resources and their performance analysed with the file accessing, resource pooling, and Job allotted by client machine with the multiple cloud providers.

In another work (Bobák, Hluchý, & Tran, 2017) addressed the problem in Multi-Cloud environments like interoperability and privacy, focussed to provide a portal offering a platform based on user requirements. The main ideas influenced the multi-cloud architecture are scalability, allows interoperability on VM, avoid vendor lock-in, create collaborative service, be independent of providers, allows flexible provider administration, and a user-friendly approach.

In another work (He & Jain, 2011), made performance attainment comparison of GoogleAPP engine and AWS. They implemented with Round-Trip Time (RTT) and network throughput with various VM provided by the two cloud service provider. This work (Qadri & Quadri, 2017) aims at education system related to cloud computing in the form of students, researchers, and faculty. All of the universities transformed their data from manual to e-based systems, which relies on scalable of data providing the costeffective secured system, highly reliable with elastic and powerful resources. This work (Duan, 2017) proposed to have a high-end future cloud computing system, from the service of virtualization and SOA brings a special challenge in performance evaluation. They evaluated the cloud service Performance attainment like service response time (delay), service throughput, service continuity, system utilization, system reliability, system enlargement, and system elasticity. Some other works (Yeasmin, Akter, Kabir, & Hossain, 2018; AL-Mukhtar & Mardan, 2014) evaluated the connectivity performance of Multi-Cloud with comparison with single cloud under various firewall situations and without firewall situations with single cloud and multi-cloud with various metrics involved in cloud networks.

Multi-Cloud computing (Akrami, Mohammadzadeh, & Vasumathi, 2014) could be an ongoing innovation and it is the use of processing services that are conveyed as a service over the internet. It has been followed by a few IT sectors since its flexible, solid, and adaptable and furthermore it licenses enterprises to maintain a strategic distance from direct infrastructure costs. One cloud provider subject could be a standard factor, at a comparable present all over access to cloud providers. In any case, these days, parallel use of cloud services given by various providers is wanted. The proficient handling of different services, needs the (Duan, 2017) analysis on performance and streamlining of all tasks in the multicloud condition. This analysis comprise of CPU scheduling, Memory usage, I/O utilization, time sharing, and money saving benefits.

The next evolutionary step of cloud services will be a Multi- Cloud services. The purpose of this work is to evaluate the network performance and analysis of Multi-Cloud systems. In this work, the authors have addressed CloudChekr, Multcloud, Odrive, and riverbed steel connect performances with different policy configuration and how far they can be utilized in user's system requirements.

3.1 CloudChekr

It is a vendor that gives a complete cloud management platform, which supports public cloud deployments in AWS, Azure, and Google Cloud Platform across the regions. It has more visibility and management control over the cloud deployments in terms of cloud computing costs, performance, and security. From this CloudChekr – AWS, they have created three idle EC2 instances across different regions. In terms of cost savings, it shows us the possible monthly savings, potential spot savings. It also produces the report for the day-to-day activity, weekly report, and monthly report of CPU utilization and gives the usage of idle resources, unused resources, and previous generation resources. From the Azure portal, default API christest_1 has been found with an average of 30 day run it gives the CPU utilization at the averages runs and peak run for a 30 days period as shown in Figure.2.

Figure 2. CPU Utilization

In Google account portal, we can create google buckets according to the Google organization policies based on the project purpose that we specify. It gives you the Google service keys, SQL servers, SSL certificates, VMs, and VPC network. After that based on the usage of the credentials that utilized, it generates the report and we evaluate how much resource is utilized for the given period of time. From this analysis, CloudChekr gives cost management of VM used. Every time this cloud management software monitors user activity by administering to implement Identity and Access Management (IAM). It also provides Resource inventory and utilization of AWS, Azure, and Google Cloud platform activities for the given period of time. Based on the deployment that gives us the report of CPU utilized, network, memory, and other resources by the VM. CloudChekr also provides some of the tools specifically for migrate, Reser ved Instance (RI) and spot management with the free version and pro plan with the pricing structure. From the price calculator, users/enterprises can decide the infrastructure they need for their requirements with a nominal price, instead of obtaining the whole infrastructure for huge prices. A sample one week utilization report for EC2 instances in CloudChekr is shown in Figure 3 and CPU instances of EC2 are shown in Figure. 4.

Figure 3. CPU Utilization Report for 1 Week

Figure 4. CPU Utilization of EC2 Instances

3.2 Multcloud

Next, Multcloud, a Multi-Cloud strategy has been used for a analysis, it comes with the deployment of many cloud providers. We work on Dropbox, Googledrive, and one drive cloud systems in one interface. This Multi-Cloud deployment gives us the flexible nature of utilizing the many clouds in a single interface. This allows us to access an overview of all file stored in one place, instead of not visiting multiple websites. Mostly, this Multcloud gives access to all the widely used cloud storage providers. It employs 256-bit AES encryption to store our files, provides direct file transfer between cloud providers, completely free and unlimited data traffic, scheduled transfer and multi-server parallel transmission. It gives free access and premium access space and security guarantee system. From this analysis, the authors observe that these benefits in terms of free plan and premium plans that are shown in Table 1.

Table 1. Comparison of Multcloud Plans

3.3 Odrive

It is a cloud aggregator offering a variety of cloud storage services. It has cloud synchronizing options and sharing capabilities, and has a universal sync folder to keep all the files in one place. It automatically unsynchronized the old files that do not access regularly. Odrive offers an additional feature to encrypt the folder and protect the data with zero-knowledge encryption, it uses AES-256 bit encryption.

In Odrive, synchronization (sync) is a great option –(Odrive, n.d) Content is everywhere. Once the file is dropped, it will automatically goes up into cloud and will be mirrored in every device and system of the user. There are two sync methods in Odrive called selective sync and progressive sync.

Selective sync is where an only certain file is locally admitted, but everything is in the cloud, where we can select the top level folders to synchronize from the web client interface. It is observed that selective sync mitigates, but kills the patient.

Progressive sync virtualizes the whole file system, so we can extend or shrink the view of all our files. The content from the local system is removed, but the files are accessible whenever it requires. We also unsync whenever the file or folder is needed to the local disk. The comparison of selective and progressive sync is shown in Table 2.

Table 2. Comparison of Selective Sync and Progressive Sync

The proposed work has focussed on Multi-Cloud computing as enterprises change their business by utilizing the intensity of the cloud, they need absolute visibility, significant insight, automation, and responsibility for their cloud speculations. CloudChekr platform creates a total image of their environment including billings, multiaccounts, resources, infrastructures, authentication, changes, and more discussed on Multi-cloud environment with ease of access and flexibility. The methodology adopted in CloudChekr is platformindependent cost analysis, thorough cloud security and a framework that had prepared for the strictest regulatory benchmarks. Or on the other hand, integrate everything with one dashboard in CloudCheckr is to access multiple CSPs across various regions provided by AWS and Google Cloud platforms. Next, the multcloud tool has been analyzed for accessing the required cloud providers under one roof and further we showed the impact of free plan and premium plan in accessing the cloud structures. Next it is analyzed with the Odrive, a Multi-Cloud management tool that holds the options of selective sync and progressive sync in accessing the file and folder structure in the cloud environment. Moreover, the aim of this paper is to adopt users to have aware of Multi-Cloud tools, access the application provided by the provider based on the user requirements in a single interface without multiple login sessions.