As a project based on a new idea, a startup involves risks, so entrepreneurs try to launch projects with minimal costs. Compared to implementing local infrastructure, cloud technologies allow to slash costs by about 25%.
Choosing a cloud provider is challenging not only because of the sheer number of options but also due to differences in pricing models, service names, and efforts to copy key market players. Here's what a startup should consider when selecting a provider for renting virtual infrastructure.
- Startup Strategy
- Cost of Cloud Services
- Cloud Types
- SLA availability
- Technical Support
- Additional Services
- Managing Virtual Infrastructure
- Reliability and Security
- Geographical Location of Servers
Startup Strategy
To understand the requirements for cloud services at the start, the startup team needs to define the project's goals and objectives. Common tasks include:
- Developing and testing new services and applications
- Data storage and backup
- Hosting websites and applications
- Providing access to a trial version of the product for partners and early customers
Once the tasks are defined, it's important to establish a budget by comparing market prices for cloud services based on pricing plans and licenses. However, accurately comparing provider prices can be challenging due to different pricing models, discounts, and variations in service offerings, particularly in their names.
Cost of Cloud Services
The cost of cloud services is determined by the amount of resources used. The more resources consumed or the more visitors to a user's site, the larger the required budget.
Typically, the cost of standard infrastructure is based on several key parameters:
- Operating system + RDP license (in case of Windows OS)
- Data center location
- Number of CPU cores
- Amount of RAM
- Storage capacity (HDD or SSD)
- Bandwidth (network throughput)
- IP address
- Additional services and licenses, such as backups and administration, if needed
When it comes to the operating system, the primary options are Linux and Windows Server. Linux is renowned for its robust security against viruses, whereas Windows Server is favored for its compatibility with a wide array of software applications. It's particularly useful for file and mail servers, often selected when a graphical interface is required. However, due to licensing costs, it tends to be more expensive.
Linux systems, on the other hand, are preferred by proficient system administration teams comfortable with command-line interfaces. They also support graphical interfaces if configured. Popular Linux distributions include Ubuntu, CentOS, and Debian, which are generally more cost-effective than Windows due to their open-source nature and lack of licensing fees.
Price: Remote Work for One Employee
On the server creation page, we set the initial parameters for the VM and plan the budget. For our example, the parameters are:
- Data center: Miami
- OS: Windows Server
- CPU: 4 cores, RAM: 8 GB
- Backup (optional)
- IPv4 address
- SSD Volume: the disk space the employee will use.
Total Expenses for the Selected Server Configuration
| Configuration Enterprise Cloud | 8GB RAM | 4 CPU | Windows 2022 Std | $28.00/mo | $0,04/hr |
| Volume SSD Lite | 40 GiB | $0.80/mo | $0,00/hr |
| Scheduled backup Mon, Wed, Fri | No. of copies 4 | $1.92/mo* | $0,00/hr* |
| Network IPv4 address | $1.00/mo | $0,00/hr |
| * The provided cost applies once the selected quantity of copies is reached. Prior to this point, the cost is dynamically calculated based on the size of the stored data at each specific hour. | ||
| Total | $31.72/mo | $0,0473/hr |
Companies often start with minimal server configuration for testing, and scale up to meet the project demands.
When choosing a cloud provider, consider the billing types:
- Allocation Pool: Payment for a dedicated pool of resources for the virtual environment, calculated monthly.
- Pay-As-You-Go: Pay only for the resources you use, charged based on usage over a specific period, such as hourly or every few minutes. This model is typical of public cloud services.
- Reservation Pool: Pay for a guaranteed amount of resources reserved for your use, regardless of availability or utilization by others. Charges are calculated monthly.
For startups, the Pay-As-You-Go model is ideal as it allows flexibility to create servers and adjust resources as needed. Payments are deducted depending on the provider, either hourly or every few minutes.
Cloud Types
There are three main types of cloud hosting: public, private, and hybrid. Each type offers different capabilities, which influence the budget for the virtual infrastructure.
Let's consider which option is most suitable for a startup project.
Public Cloud
In a public cloud solution, the provider owns and manages the virtual IT infrastructure and rents it out to users. The provider consolidates a certain amount of computing resources into clusters. Clients can use these resources to create virtual machines with the necessary specifications until the resources are exhausted. Providers typically have monitoring systems that notify users in advance when resources are running low, prompting the provider to acquire additional equipment.
For startups, a public cloud is an option. It requires minimal investment, as there is no need to purchase equipment - the physical infrastructure is managed by the provider - and no need to hire staff for maintenance and support. This type of cloud can be deployed in minutes.
A public cloud allows for rapid scaling up to the limits specified by the provider. This is particularly relevant for e-commerce projects, where there is a high load on services during peak sales periods. As the project grows, the startup gains more clients, increasing the load on services.
The limitations of the public cloud include:
- The larger the project, the more computing resources are required. Not all providers can meet this demand.
- Many clients are concerned about data privacy, so it's crucial to ensure compliance with the laws of the country where the provider's servers are located and where the company is renting them.
- Users are entirely dependent on the provider. In case of failures, users cannot influence the speed of problem resolution. However, you can be assured that professionals are handling the issue.
Private Cloud
A private cloud is a solution where the virtual infrastructure is owned by a single company. This type of cloud can be organized independently using the company's own resources or by renting equipment.
Unlike the public cloud, the user does not need to share resources with other clients, so access is exclusive, resulting in better data protection.
With a direct connection to the private cloud, the load can be distributed between physical and virtual infrastructure. Users can reallocate resources among different departments and branches within the company, depending on current needs.
For a startup, one of the main disadvantages of a private cloud can be its cost. The price depends on the custom cloud project created by provider specialists according to the company's requirements. This solution is typically chosen by projects with substantial initial capital and a strong belief that their project will grow and pay off.
It's also important to note that, unlike a public cloud, quickly creating virtual infrastructure is not possible. Deploying and configuring a system tailored to the company takes time.
Scaling is possible within the allocated cloud, which requires the ability to plan computing resource consumption in advance.
Hybrid Cloud
A solution where a private cloud is integrated with a public one, allowing to leverage the advantages of both cloud types.
Additionally, there's also the concept of multicloud, where clients can use services from different providers. Technically, this is even more complex and costly, hence not the most suitable option for startups at the beginning of their journey.
Hybrid clouds hold the leading position on the market, however, startups typically opt for public cloud due to its cost-effectiveness and quick deployment.
SLA availability
Service Level Agreement confirms the availability of infrastructure and prompt response to incidents. This agreement between the provider and client specifies guarantees and parameters of cloud service availability:
- Network availability;
- Storage Area Network (SAN) performance;
- Technical support conditions;
- Scheduled maintenance;
- Financial guarantees.
Within the SLA, all metrics are defined in terms of timeframes for resolving specific issues. It also outlines fixed time intervals for addressing non-standard client requests.
Cloud providers bear financial responsibility for SLA breaches, as stipulated in the document.
Technical Support
The work of cloud provider's technical support is one of the guarantees of infrastructure availability. The service handles user requests, notifies about incidents, and is responsible for their resolution. The format of operation and processing time for user requests are documented in the SLA.
Here are the main criteria in the SLA that technical support of the provider must comply with:
- Round-the-clock daily operation;
- Access via various communication channels: phone, email, ticketing system, website chat;
- Incident response time specified in the SLA;
- Assistance in resolving non-standard requests;
- Timely notification of scheduled maintenance and incident resolution.
Before deploying infrastructure, we recommend testing the technical support: ask a question and see how quickly specialists respond. Such a simple test will help save time in case of a real incident.
Additional Services
In addition to virtual servers, cloud providers offer clients additional services: Remote Desktop (RDP), Managed Kubernetes, DNS (Domain Name System), CDN (Content Delivery Network), and others.
On one hand, this allows businesses to scale. On the other hand, startups rarely need a large portfolio of services, especially at the beginning.
Therefore, before deploying infrastructure, it is important to determine exactly what a startup needs to launch its project. For example, for product development and testing, it may be sufficient to:
- Deploy virtual servers and remove them later if necessary. You can start with three - for development, testing, and production.
- Configure an isolated network.
- Install internal software: email, CRM, document management, and accounting.
- Set up a web server and host a website.
- Install necessary software.
To simplify software deployment, providers offer application marketplaces where you can choose what to install when creating a server. For instance, many offer to create a server with a pre-deployed WordPress system, which allows creating and managing website content.
Managing Virtual Infrastructure
The aforementioned services are ordered through the provider's control panel. In a small startup team where members may not have experience with system administration, they often have to figure out how to manage the infrastructure themselves. Therefore, it's important to choose a provider whose control panel is as user-friendly as possible.
Providers geared towards solutions for large clients with complex IT infrastructure often integrate many features into their panel, making it difficult for non-specialists to navigate.
Key functionalities that should be in the control panel:
- Deploying Virtual Servers
- Choosing operating system templates - such as Windows Server, Linux distributions, FreeBSD.
- Ability to modify server configuration.
- Uploading custom ISO images - files containing an exact copy or archive of the disk storing data and OS.
- Server rebooting.
- Remote access to virtual servers.
- Availability of backups.
- Creating snapshots of data - file system, server. This allows restoring the virtual machine to a working state in case of failure.
- Monitoring server load statistics.
Also, in the personal account, there should be the following options:
- Various payment methods: credit cards, electronic wallets, and for legal entities, payment by invoice.
- History of financial transactions conducted from the personal account.
- Management of account settings: changing login, password, and email.
- Updating account owner information and legal entity details.
- Contact with technical support.
- Adding participants for collaborative work.
- Let's consider several examples of control panels.
Reliability and Security
Every business's data must be protected, and the provider should hold appropriate licenses and certificates. Additionally, providers must keep equipment in reliable data centers that operate without interruptions. Certificates and standards guarantee the data center's reliability and security, such as compliance with operational and protective requirements for data storage.
All infrastructure resources have backups, such as multiple power and cooling supply channels. During maintenance windows (typically 8-12 hours), only one of these elements remains active.
There exist criteria and standards, like PCI DSS for credit card data protection or ISO standards for service processes: ISO-9000 for general quality management, ISO-27001 for information security, and ISO-14000 for environmental aspects.
Despite startup needs differing from those of larger organizations initially, it's advisable to choose a data center that meets high service standards.
Geographical Location of Servers
On the internet, latency is determined by the distance the signal travels from the server to the recipient.
The physical distance between data centers and clients directly impacts the speed of applications, websites, and services.
To leverage that, a cloud provider can place equipment both in leased racks at different data processing centers and in their own data centers. There is no noticeable difference for the client between these options.
If a startup is launching a project abroad or plans to expand territorially, it's advisable to choose a provider with geographically distributed locations. For example, if starting a startup in Europe, it's better to find a provider with a data center located in the country of launch or in the nearest region. Data centers on other continents would not be suitable in this case.
Ready to ask a question? Let our technical experts help: Use our online chat to get a custom solution for your case.
