Gartner predicts that in 2022, spending on IaaS cloud infrastructure will have grown by 24% compared to last year and will reach $106.8 billion. Interest in the cloud fueled by numerous cases in which businesses were able to optimize IT infrastructure costs, make it flexible and scalable, better prepared for bursts of load. These are the results not so much of migration to the cloud, as of the right choice of a service provider.
Virtual infrastructure costs are rising, but companies are looking for affordable data hosting options and do not want to increase spending on IT infrastructure deployment and support. We suggest assessing the reliability of cloud providers by seven criteria first, and only then compare prices.
Reliability
The continuity of business processes, financial results and even a company’s reputation often depend on whether it can secure data and provide uninterrupted access to it. This depends on the parameters of the data center your provider uses.
To assess reliability (Tier) the classification from the Uptime Institute standard is used, where data centers are divided into four groups, depending on ways of backing data up.
- Tier I – Basic infrastructure without redundancy (N). If one component fails, data is lost. Tier I data centers are unreliable and irrelevant.
- Tier II is an infrastructure with minimal redundancy of utilities, where additional elements in air conditioning and uninterruptible power systems are used (N+1). With little redundancy, data center customers are less affected by equipment failures, but in case of serious errors data can be lost, and for preventive maintenance it will be necessary to shut down all systems.
- Tier III is an infrastructure where, thanks to dual redundancy (2N), routine maintenance can be performed without shutting down the load. All engineering components and power lines are duplicated. If something breaks down or routine infrastructure maintenance is needed, customer data remains available.
Tier IV is a completely fault-tolerant infrastructure (2N+1). Engineering systems are not only redundant, but also contain one additional component each, so the probability of data loss is close to zero.
The classification developed by Uptime Institute is not the only one. Moreover, 90% of European data centers are not certified for compliance with Tier III and Tier IV criteria, although in fact they meet them. Since 2016, the EU has adopted a new standard for assessing data center reliability, which is confirmed by the certificate EN 50600, which sites with at least N+1 redundancy, complies with the international standard ISO 27001 for information security and quality management standards ISO 9001.
Colobridge customers’ data is stored in data centers of the world’s best data center operators Equinix and KDDI Telehouse, located in the European IT capital – the German city of Frankfurt am Main. Their engineering infrastructures and security systems comply with Tier III+ standards (Tier IV in some containment areas), and the availability guarantee is over 99.99%.
Services
Pay attention to the functionality offered by the provider: in addition to the number of processor cores, memory size and storage capacity, it is important to have SLA on the technical characteristics of virtual machines, the number of input/output operations per second and the bandwidth of network ports.
Backup and emergency recovery tools, DNS hosting, public APIs, network attack protection, 24/7 service availability monitoring systems will be helpful. If reliability is a priority, choose a cloud provider with two or more geographically distributed data centers to reduce risks. If you’re not ready to give up your servers completely, explore the prospects and benefits of a hybrid cloud. And if your applications collect and store large amounts of unstructured data, consider contracting with a provider that has object storage.
Security
Find out how your infrastructure will be protected at the physical and software layers. In IaaS, data protection falls on both the customers and the provider. The latter does not create virtual machines, but rather provides virtual capacity for deploying customer business applications. Accordingly, the provider does not control the data itself, but protects the virtualization environment and administrator workstations, ensures the functioning of antivirus software and the updating of antivirus databases, and regulates the collection and storage of logs. At the cloud level, the provider is also responsible for software updates and vulnerability detection and mitigation. At the physical layer, the provider backs up the main nodes of its infrastructure, restricts access to the data center, and installs security alarms and video surveillance cameras.
An additional advantage would be the location of the provider’s facilities in a large European data center (Germany, the Netherlands) or in the United States. The good news is that in many Western countries it is only possible to remove equipment from the data center by court decision. However, when choosing the right jurisdiction, consider that the greater distance between customers and the physical data center increases delays, which may be critical for some applications – for example, online games, streaming services, delivery services.
Some other significant security guarantees include: absence of a single point of failure in the data center, 24/7 monitoring tools, international ISO/EIC certifications, explicit SLA guarantees for power, cooling and data transfer channels (ideally 99.99% or higher).
Colobridge’s high level of corporate service is confirmed by international certificates ISO 9001: 2008 and ISO/IEC 27001:2005.
Payment
Consider the available payment options and their interchangeability – it’s good if there is a choice. Familiarize yourself with the requirements for paying bills: sometimes money can be transferred in both foreign and local currencies.
Many cloud providers already use the Pay-as-You-Go model, where customers pay only for the amount of capacity actually consumed. This is convenient and fair in contrast to paying for a fixed amount of resources, when it does not matter whether you use them in full or not.
Testing
During the free test period, you will learn how the cloud copes with various loads, talk to support specialists and estimate costs. It is not always possible to test the service qualitatively in two or three days, but it is feasible in a week. Therefore, priority is given to those providers who allow you to test the cloud infrastructure for at least seven days and are willing to extend this period if the client needs a little more time.
Support
You can always rely on basic technical support over the phone, but how good it is and in what mode – it’s better to find out from external sources, such as customer reviews. It is important that you get help 24/7, in your native language and through several communication channels: by phone, by e-mail or via a feedback form. Some providers create an extensive database of ready-to-use questions and answers where you can get information without contacting support, or run chatbots in popular messengers that respond to problems faster than live operators.
Find out what additional services are available from your chosen provider. Perhaps their specialists will completely transfer your data to the cloud and will be able to administer your infrastructure in the future. Find out the cost of such services – it may be cheaper than hiring an engineer.
Customers
The list of public customers and their successful cases determine the level of trust in the cloud provider. If there are market leaders on the list, you can use IaaS infrastructure on the same facilities.
Adobe, IBM, Yahoo, Amazon, Google and Microsoft cooperate with data centers that host virtual machines for Colobridge clients. Companies such as Multiplex, Glovo, Softlist, Deca, Astor and Stalkanat cooperate with Colobridge itself.
Get more information on ‘Infrastructure as a Service’ or ask our expert questions to make your final decision on choosing a cloud provider.
