
Until now, not everyone understands what
IaaS means, and those who understand the meaning of this term come across a large number of companies that provide virtual server rental services, but hide significantly different solutions and technologies behind the IaaS abbreviation.
Why is it important to single out a separate segment of the so-called “corporate IaaS” and understand its differences from the massive “unincorporated” market of renting cloud servers? Let's try to answer this question.
')
The trend of “cloud” orientation is gaining quite serious momentum, and the cloud infrastructure as a service (or IaaS mentioned earlier) is announced as the most popular cloud service in the corporate segment.
And what do we see in the rental market for cloud servers?
All IaaS service providers as one declare guarantees of the availability of cloud services 99.9% per month or more, mobility and scalability, cost reduction, professional administration, data security, reliable backup system, flexible technical support 24/7/365 with all the consequences financial responsibility in case of unavailability of service. It sounds more than optimistic, but at the same time the cost of the service called “IaaS” may vary by different suppliers.
What caused such a difference in cost?
And the difference is - as usual - in the implementation details. On which site the hardware will be located, what the hardware will be, which virtualization platform will be used and much more will depend not only on the cost of the virtual servers, but also on the presence or absence of additional capabilities for working with these servers.
On the other hand, let's say, you face the task of renting several virtual servers with a given configuration. When buying such a service, some customers wonder where their data will be physically stored and how the virtualization infrastructure will be implemented. To others this information will seem perfect unnecessary. And both will be right, because for some tasks (let's call them corporate) this is really important, and for others, “unincorporated”, the primary criteria are price and usability.
It turns out that we have two IaaS market segments - corporate and non-corporate (or mass). To begin with, we will analyze these segments from the suppliers of both IaaS implementation options.
Mass IaaS Provider (vps / vds hoster)
As a rule, such providers are traditional web-hosters, expanding their package of services with a hosting service of virtual machines.
The implementation of the cloud platform in this case is focused on automating self-service and minimizing the cost of placement, the hardware component and the hypervisor. These are mostly private server rooms (with reliability not exceeding Tier 2) inexpensive servers with hard drives on board, deployed on top of them Open Source (free, open source) or container virtualization, plus a control panel (typical boxed, or in-house development).
At the output, we obtain satisfactory reliability (with the regular backup option enabled), somewhat limited functionality, relatively humane cost, and a convenient management interface, mostly with full self-service, without the involvement of hoster specialists. In such a model, the hoster builds up a familiar and cost-effective streaming process of mass service of a large number of mostly small subscribers, and subscribers receive essentially “virtual machines” cheaply and angrily.
Corporate IaaS Provider (IaaS Service Provider)
Large data centers and system integrators are gradually becoming suppliers of corporate IaaS. Separately, we can distinguish the so-called “cloud integrators” that appeared on the wave of cloud technologies, which have absorbed the strengths of the first two industries and do not pull their vestiges.
One way or another, such companies do not operate in the mass market, being targeted at the corporate segment. Therefore, the cloud platform is located in reliable data centers with a Tier 3 reliability category, and the storage system is included in the hardware. In this case, all critical nodes of the hardware and network parts are duplicated. Generally, proprietary corporate-level solutions are taken as the hypervisor, and the control panels are full-time control panels of these hypervisors or there is no cloud management interface at all (when all settings are made by the provider on request).
As a result, we get a relatively more expensive solution than the first one, which is less convenient in terms of automated self-service capabilities, but with much more flexible functionality, high level of reliability and fault tolerance. Here we are not talking about virtual machines, but about a virtual infrastructure or a virtual data center, which require fine-tuning to the needs of a large client and integration with its existing infrastructure.
Who and why is ready to overpay for a more heavy solution?
For our convenience, we divide consumers into two conditional groups:
Companies using large and / or critical information systems
By a large information system, we will mean a service that serves a large number of users - employees of a company, its clients or partners. The more such users, the greater the damage from its downtime or disruption.
On the other hand, a critical information system is a corporate application on which the core business processes of a company depend. Failure of a critical service key business process can halt all the activities of an organization and lead to irreparable losses, up to the complete cessation of the company's activities.
The important point here is the degree of IT penetration into the company's business processes. The higher the level of informatization, the greater the dependence of business results on the stability of the main IT services.
A typical example is a large retail network with a centralized IT infrastructure, a large online store, a federal logistics company using a single automation system, etc. Any simple corporate information system for such companies leads to substantial losses, and in some cases may even threaten a complete cessation of activity.
The main tasks solved by such companies in the IaaS cloud:- the removal of the entire server infrastructure to the cloud;
- deployment of corporate information systems in the cloud;
- deploying B2C and B2B sites to interact with customers and partners (loaded online stores, marketplaces, etc.);
- reserve site (DRS) .
Companies that do not use large and critical information systems in their work
On the contrary, if the degree of penetration into IT is low, the scale of the business is small, then the dependence on information systems is significantly lower. For example, a small car dealership or engineering company, where the majority of employees work mainly with local applications. In a situation where the main server has failed and it takes several hours to restore it from a backup, such companies will not have serious problems. Similarly, a small group of software developers will calmly survive even the many hours of its simple server infrastructure dedicated to the development environment.
The main clients of mass IaaS are placed on virtual machines:- typical sites with a small load,
- terminal servers with 1C for several users
- development environment for a small group of developers.
It turns out that, depending on the size and severity of its application, different requirements for the following characteristics may apply:- reliability,
- fault tolerance and redundancy,
- Information Security.
In a situation where the risks and size of potential damage from IT failures increase with the growth of informatization, the increase in the cost of computing resources becomes justified. But interdependent categories of cost and reliability - this is just one of the aspects on which the concepts of corporate and mass IaaS clouds differ. Below we present the main characteristic differences between these segments of the virtual infrastructure market.
Character traits | Corporate IaaS / virtual infrastructure rent / virtual data center rental | Unincorporated, massive IaaS / VPS hosting / VDS hosting |
---|
Client tasks | Corporate IT infrastructure. Reserve site. Loaded online resources. | Small software development projects. Collaboration with the accounting system of several users. Hosting sites. |
Priorities customer | Reliability, fault tolerance, performance, security, flexibility, technical support. | Cost, launch speed, simplicity and convenience of the interface, convenience of payment |
Supplier Priorities | Reliability, availability, performance, isolation, security, flexibility, functionality | Reducing the cost of resources and support, simplifying functionality, increasing convenience, automating self-service |
Site reliability level | Tier 3 and up | up to tier 3 |
Equipment | Reliable server + enterprise storage. Using Enterprise solutions from leading manufacturers. | Budget servers with disks on board or using entry-level storage systems. Using low end or middle range solutions |
Type of virtualization | VMware, less Microsoft Hyper-V | OpenVZ, Parallels Virtuozzo, Parallels Cloud Server, KVM, XEN, FreeBSD Jail, Microsoft Hyper-V |
Reliability assurance | Duplication of the main hardware nodes. High availability feature of the hypervisor. Backup | Backup. Fault tolerance at the level of the customer’s application |
Security | Using effective methods and security tools, following in step with the times, meeting the requirements of security standards (for example, PCI DSS). Protection against all kinds of attacks, including DoS, DDoS. The ability to implement two-factor authentication and other means of protection | The implementation of security at the minimum level. Optionally - DDoS protection |
Typical secure connection options | Organization of software VPN. Hardware VPN. Connecting via a dedicated channel - directly or through a traffic exchange point. | Organization of software VPN |
Self catering | Part of the operations is implemented through contacting technical support. | Maximum self-service via management interface, interaction with support only for incidents / |
Management interface | Sophisticated, great functionality for flexible customization | Simple, convenient, functionality is limited. |
Load Management in a Virtualization Environment | Load balancing virtualized infrastructure by means of the scheduler. Providing automatic movement of virtual machines in accordance with the calculations of monitoring services | Lack of automated load balancing mechanisms for virtualized infrastructure. Moving virtual machines manually. Human factor |
Contract | Signed and agreed individually | Contract offer |
Payment | Postpay. Cashless payments, accounts with VAT, acts. | Prepay. Payment by bank cards and electronic money |
Tariffs | Individually agreed resource value Dependence of the cost of resources of a virtual data center on the quantity and quality of resources, as well as on the model of using a virtual infrastructure | Fixed tariff plans with included resources Typical tariffs with a limited set of offers |
Additional and higher services (examples) | Software rental OS Administration Monitoring SaaS services (Exchange, 1C, ...) | Domain sale Hosting sites Mail hosting File storage |
Service system | Business processes are sharpened by taking into account individual quality parameters for each client with their own unique level of service delivery (SLA) Individual approach to every customer The possibility of deviation from the framework of the contract concluded with the client Solving emerging non-standard tasks, with the possibility of allocating additional resources Willingness to solve very complex tasks. Qualified professionals, even on the first line of support | The business process is put on stream, the main task is the mass service of a large number of typical clients. Lack of an individual approach to the customer, all within the framework of a standard contract Lack of flexibility in service level agreement Technical support solves exclusively standard set of tasks. Relatively lower qualification of support specialists |
Both considered IaaS options can be equally useful and interesting for different categories of clients and customers. In this context, it turns out that an unincorporated IaaS versus a corporate one is to rent a virtual server vs to rent a full-fledged virtual data center.
In which case, the choice of corporate or unincorporated IaaS would be justified?
In which case, the choice of corporate or unincorporated IaaS would be justified? Suppose the developers of a certain company turned to the cloud provider, the IT penetration rate is rather low. The company's developers are working on a temporary project, which is why it became necessary to rent cloud services on a temporary basis. As the saying goes: "finished the project and forgot." Availability of data for this company is important, but not critical. Developers need to quickly and cheaply gain access to dedicated resources. At the same time, the company is not faced with the task of managing a large project; everything is rather primitive and straightforward. In this situation, it is more logical to pay attention to the unincorporated IaaS, since the project is temporary, not very expensive in terms of resources, does not require Enterprise solutions to solve the task, and you can save money. Similarly, when choosing an IaaS provider for hosting a standard site or a small online store, you should focus on mass hosters, since the cost of accommodation and convenience of work will be of paramount importance.
Or another example: a large company-developer needed a highly accessible, high-performance cloud platform to support the process of streaming testing of new versions of a wide range of subsystems. The customer wants to see adequate automation of work with the cloud. Availability, stability, the possibility of flexible integration and unlimited increase in the resources of the cloud component, as well as flexible payment mechanisms upon the use of resources - will be important criteria for choosing a platform for such a task. And this is a choice towards corporate IaaS.
And if the existing IT infrastructure in your company with a large number of clients, domains, terminal servers, heavily loaded application servers and databases requires a large-scale upgrade, and a decision has been made to move the servers to the cloud of the corporate IaaS provider, then this is the right way of thinking.
Thus, the main criterion for the choice of two types of IaaS is the availability of an adequate assessment of the level of IT penetration in the company and a clear vision of the answer to the question: will there be enough opportunities available to one or another hosting service provider? Will it be able to ensure the operation of my company's IT systems, when transferred to the cloud, taking into account the characteristic features of the current IT infrastructure?
When transferring infrastructure to the cloud, you need to analyze a lot. If the client company has growth potential and an understanding that the resources allocated today may not be enough tomorrow, consider all the risks. Will a cloud provider offering IaaS as a service cope with the demands of increasing resources and workloads that may follow in the near foreseeable future? Does the cloud provider offer a so-called “On-demand self-service”, namely the availability of services in any volume, if there is a need for the customer? Do not miss the issue of security, both physical and network. Most of the suppliers, as a rule, far from the corporate format, do not provide any exact information about the principle of operation of their systems, and therefore about the quality of the protection they provide.
Here we looked at two polar models of providing IaaS services with their distinctive features. Of course, there are mixed and intermediate versions of varying degrees of "corporate", reliability, flexibility, etc., etc. For more information about the corporate version of the provision of IaaS services, we write in the
first blog about corporate IaaS . As an illustration of a typical corporate IaaS provider, you can cite
IT-GRAD , and as an unincorporated
1cloud.ru