How to Choose a Reliable Data Center Provider: 8 Success Factors

Started by Slip, Jul 30, 2022, 03:43 AM

Previous topic - Next topic

SlipTopic starter

A reliable data center provides uninterrupted power and cooling for installed apparatus, uninterrupted operation of your servers, communication channels and multi-layered protection to ensure the safety of critical data.
But not all data centers are created equal. Very often, the reliability of a data center and the quality of assistance  depend on the company that owns it.





Some data centers (especially overseas) are owned by real estate companies, others by holding companies that own other types of high-tech or non-high-tech businesses. Often, hosting companies have only minimal experience in owning and operating a data center. It is better to choose a reliable, experienced provider that specializes in data centers.


Obviously, the cheapest option does not always mean the best assistance for your money. If you don't want to sacrifice IT security or risk critical data, then it's best to avoid providers who consider the data center to be "just another high-tech warehouse", lack experience, or rent the site from a different owner.


There are some critical success factors to consider when choosing a data center. Among them:

     Company experience.
     Her financial stability.
     physical infrastructure.
     Physical security.
     Customer assistance and support.
     Service Level Agreements
     Suggested tariff plans.


company experience

Undoubtedly, the experience of the company - the owner of the data center, how long it has been in the industry is of great importance. Keep the long term in mind: choose a data center provider that can host your IT equipment for years to come. Look for a data center provider that has been known in this business for more than one year (or better, 10 years), which indirectly confirms the reliability of the company.

A significant track record also shows that the company has accumulated enough experience in operating a data center to provide a decent assistance  and high reliability of the operation of its website.

If the provider has additional technology competencies, it is useful to find out how closely they are directly related to the data center. Are the services of renting space in the data center or accommodating customer apparatus (colocation or collocation) accompanied by other services, what services are provided? Is there a dedicated division dedicated to providing the best customer assistance ?

Does the provider have one or more data centers? The more sites (including foreign ones - rented or own), the more serious the business. It's good if the provider is a growing, developing company, that is, it builds or acquires new data centers, and not just operates existing ones. This shows that the company is investing in the interest and introducing new technologies.

Ask the provider for a list of clients, if it is open information, get acquainted with "success stories" and customer reviews. If large and well-known companies and organizations are among the customers, then they have already done due diligence, studied the issue and are ready to trust this provider with their IT infrastructure and critical data.

Ask your ISP if they serve other companies in your industry. If a data center owner specializes in serving certain industries or vertical markets (for example, healthcare, finance, or oil and gas), then he is likely to be well aware of the needs of customers in these industries (for example, understand the requirements for personal data storage and can act as the operator of such data).

It also doesn't hurt to go to the provider's website for management biographies, if there is such knowledge. How long have they been with the company? What experience do they have in the data center industry? If executives have rich data center experience, they are likely to have a better understanding of how successful data centers work, as well as the needs of their customers.

Finally, ask if the vendor has any strategic partnerships that could be beneficial to your company (for example, they might partner with a SaaS, PaaS, or IaaS cloud service provider, a hosted server platform provider, or a security solution provider). Also ask if the provider has received any industry awards for customer assistance .


financial stability

You need to make sure that the owner of the data center has the assets to provide long-term services. It is risky to deal with a provider that is struggling to meet its financial obligations. You don't want to choose a provider that could go bankrupt in a few years and leave you in a difficult situation.

If the provider is a public company, then its latest financial indicators are available in annual reports, published in the media. Private companies are usually less willing to share financial data and it may take a bit more effort to get this information. State structures are subject to more thorough financial scrutiny.

You can ask the provider for the latest financial reports. For foreign companies, financial performance is usually checked by an independent auditor, such as Ernst & Young or KPMG. The key element is transparency. If a supplier is unwilling to share financial information or conduct an independent audit, this should be cause for concern.

If you manage to get the relevant financial statements, ask your accountant to analyze them in detail from the point of view of fulfilling short-term and long-term financial obligations.

According to foreign experts, it is also worth studying the following elements:

    Does the provider provide insurance for customers, for which insured events? Who is an insurance company?
    It is good if the provider has a revolving line of credit from reputable banks. A public company may have a good credit rating from a reputable agency such as Standard & Poor's or Moody's.
    The supplier must have access to bank loans or investors to support future growth.


An important issue for large customers to consider when considering a data center provider is how much of the business you will provide to them. If hosting your IT apparatus will provide 50% of his interest, he may give you special conditions. But if you decide to double your IT fleet, you may be overreaching the provider.

▍Risk assessment

Before considering other critical factors, let's focus on the risks. The #1 question for every company choosing a data center provider is "How secure will this be for my IT infrastructure and mission critical data?". For each provider under consideration, the risks of server failure or data loss need to be assessed.

There are three main areas of risk assessment:

    Risk of loss of servers and/or critical data due to natural or man-made disasters (hurricanes, floods, earthquakes, fires).
    Risk of data loss due to data center equipment failure (power outage).
    Risk of data loss due to security breach.


To assess these risks, you need to analyze the location, physical infrastructure, and security systems of the data center where you plan to host your servers. The data center owner must have a disaster recovery/business continuity plan. Each data center should have specific procedures that describe what to do in the event of a power outage, natural disaster, etc. The provider should have plans in place to ensure the security of customer data and apparatus, as well as the IT infrastructure at each website.

These plans should be aligned with best practices such as international business continuity standards. Data center employees should be well trained in disaster recovery procedures. In addition, the facility should regularly check its failure prevention systems (i.e. automatic power transfer, fire protection systems) to ensure they are working properly in the event of an incident.


physical infrastructure

Take a very close look at the physical infrastructure of the data center where you want to host your servers. Ask the owner of the data center to give you a tour of the facility for your company.

▍Location

No data center can be completely protected from natural or man-made disasters. Wherever a website is located, it is vulnerable to some type of disaster, be it an earthquake, tornado, hurricane, flood, fire, etc. However, reliable data centers are built to withstand almost any disaster.

Many infrastructure elements help achieve this, but the foundation is the actual physical location of the data center. For example, the data center should be built on a hill where it is inaccessible to flood waters. Data centers should not be built near airports or in close proximity to any industrial facilities (nuclear or chemical plants).

The data center building can be a reinforced concrete structure with special characteristics designed for different types of natural disasters that are typical for the area. Sometimes the data center is located in former bomb shelters. A data center in an earthquake-prone area needs to be earthquake-resistant to withstand a sufficiently strong earthquake. A data center in an area prone to hurricanes or tornadoes must have storm protection.

It is also very important that every data center has a reliable fire protection system such as a smoke detection system (VESDA).

▍Redundancy of power supply and cooling systems

A reliable data center has several levels of redundancy. Its power supplies must be fault-tolerant so that if one power supply fails, it will automatically switch to other power supplies. These may include:

    Backup input from another substation. If possible, the data center should have two independent power inputs. If one fails, the other will continue to provide power to the data center.

    Diesel generators. Many sites have a DGU that provides power to the facility in the event of a local power failure. Diesel generator sets with sufficient fuel supply can keep the facility operational for several days, and even longer if fuel is delivered.

    Uninterruptible power supplies (UPS). Many data centers have a room filled with battery shelves. If the main power source fails, the UPS supplies power to the website until the backup power sources are connected - the standby input or the generator set.

There are also a number of factors to consider regarding the physical infrastructure of a data center that are not related to disaster preparedness. Often this refers to your interest, your current IT infrastructure and your plans.

For example, choosing an actual data center location usually depends on your company's needs. Some companies prefer to have their IT infrastructure within reach of the company's headquarters or its core business.

Other companies do not experience problems with the deployment of IT infrastructure in a remote city. If only there was an airport so that they could fly in and check their computer facilities. This is usually best suited for companies with multiple regional branches and a geographically dispersed structure.

▍Available data center space and future expansion

You need to have an idea of how much space, power and cooling capacity your IT assets will need, and if the data center has that amount of resources. A good data center has space and capacity reserves for future growth. All you need to do is evaluate your future IT infrastructure scaling needs and make sure the data center is able to meet them. Ask how many watts per rack are available, what is the maximum/minimum power density.

Ask for a rack layout and data center plan. Use it to determine how much space you need for your current need and how much additional space is available in case an expansion is needed.

▍Cooling specifications

In addition to redundant power, a reliable data center keeps your IT equipment cool. It must meet or exceed ASHRAE temperature and humidity standards.

▍Communication channels

A reliable data center offers highly reliable and multi-carrier communication channels with the required bandwidth. Sometimes there is a payment depending on the bandwidth used, you can flexibly scale the bandwidth during periods of maximum and minimum activity.

In some cases, it becomes necessary to create disaster-proof configurations with replication between remote sites or to organize backup to a remote data center.


physical security

    A trusted data center has multiple layers of security to protect your IT infrastructure and data. Visitors to the facility will need to pass through several security zones. Here is an example of such layered security:
    The data center building is surrounded by a high fence. Guests must register at the entrance to the car park.
    The entrance area is a protected area, separated from the rest of the facility and monitored around the clock by the security assistance.
    Visitors are required to present a photo ID at the entrance, sometimes a biometric scan (fingerprint and retina scan).
    Visitors must be escorted to their counters by security personnel.
    In some cases, the client's IT infrastructure is in a closed "cage". Visitors must have a key, an RFID key card and/or a PIN to enter this area.

The data center should have established security procedures that govern who has access to the facility, how visitor passes are issued and registered, and procedures to be followed in the event of a security breach. The data center must be manned by security personnel working 24/7.

    Education. Security personnel must be well trained and thoroughly familiar with website security procedures.
    Checks. Employees must be thoroughly screened when they are hired.
    Non Disclosure Agreements (NDA). You need to know what types of NDA security officers are required to sign.
    Subcontractors. If the data center owner uses third party companies for security, you need to know who these subcontractors are, how they screen their employees.

Finally, the data center must meet the regulatory requirements that are important to you. Ask the supplier for documentation that the data center has passed such a check (abroad, these are the requirements of HIPAA, SSAE-16, SOX, FISMA, PCI-DSS, etc.).


customer assistance and support

A reliable data center provides its customers with high quality assistance and support. It should be manned by well-trained and attentive staff on website 24 hours a day (in addition to security personnel). Site personnel will be responsible for managing and monitoring the condition of the facility and its systems, maintenance and storage, handling and delivery, etc.

The control center controls all data center systems, including IT systems, power levels, temperature, humidity, video surveillance. Data center customers need to have access to real-time reports to constantly monitor the health of their information systems.

In addition, the data center must provide a set of established procedures for daily operations, routine preventive maintenance, and periodic system testing. Employees should be trained in established procedures to determine what to do in the event of an "event" (such as a power outage).

A system is also needed to notify customers of such a situation. Data center staff should provide you with support when you move, install apparatus, or expand your space.


Service Level Agreements

Choose a provider that offers a assistance level agreement (SLA) with a written guarantee of uninterrupted power and uptime (such as 99.999% or five nines). The SLA should also state how much compensation you will receive if the supplier fails to meet its obligations.


Suggested tariff plans

Find out what basic hosting plans are offered and how the plans are structured. Questions you should ask include:

     Is there a monthly/yearly fee?
     How will the price of expanding services be affected over the life of the contract?
     Are there any hidden fees?
     In the case of a lease, look at its terms and conditions: are there any unexpected expenses, such as property taxes, utility bills.


conclusions

Your goal is to select a data center provider that is a reliable business partner for you. Your company's success depends on the high availability of your IT environment and the security of your critical data. Thus, one should carefully study the provider before signing the agreement, requiring as much transparency as possible. The more knowledge you get, the easier it will be for you to make an informed decision and choose a reliable partner for the long term. Reliable Data Center Service Provider is an experienced, stable and growing company with highly reliable and well-designed sites.

But when is it worth renting space and / or IT resources in a commercial data center, and when is it worth building or expanding your website? Finally, let's touch on this topic.

Build or rent?

Today, data is a critical asset for almost any interest, and the need to host and maintain the ever-growing "digital assets" is increasing exponentially. When an organization is faced with the need to expand or modernize its data center, it faces a choice: build a new data center, refurbish an existing data center, or find a partner to host its IT apparatus. Choosing the right assistance provider can lower your total cost of ownership, converting capital expenditures into operating expenses.

Why is there a need to expand data centers at all? Here are just a few of the possible reasons:

    Increase existing storage capacity. Accumulation of more data.
    Equipment obsolescence. Technology develops exponentially, the "life cycle" of IT apparatus is only a few years.
    Business expansion. The need to increase resources while maintaining the level of assistance.
    Market trends. Support for corporate mobility and Internet technologies (IoT).
    Mergers and acquisitions. Combining disparate assets into a single structure.
    Separation of businesses. The success of one direction or another or the need for optimization leads to the separation of a part of the company into an independent interest.
    Transition to big data and analytics. Use of collected and accumulated data in the interests of business.
    Emergency recovery. Protecting "digital assets" and ensuring business processes.


Uncertainty of the future

When planning to upgrade existing data centers or build new ones, identifying future needs can be difficult, and significant investments typically take a long time to pay off. The following trends should also be taken into account:

    The workload is gradually moving to the public cloud. What will your opportunities be in 5 or 10 years?

    Technology is constantly changing. What applications will be vital to your business in the future?

    The need for computing power and storage resources is not constant. Can you scale it up or down dynamically with the available staff, budget, and capacity?

    Time to market is more important than ever. There is a growing need for new applications and workloads. How soon will you need additional capacity and power?


Every company's data center needs are fluid and changing, and you need to think carefully before you start building your own website, a capital expenditure that includes facilities, maintenance, and power.

CAPEX to OPEX

If data centers are not the backbone of your business, don't dedicate so many resources – financial and human – to them. Partnering with a trusted data center provider will free up time and capital that could be better spent growing your interest. As the cloud model spreads, the need for in-house data centers will decrease. Why create a data center that can be significantly underloaded in a few years?

Providers have already invested in large data centers around the world. Take advantage of "economy of scale" and direct your capital to business development. The provider already has the necessary infrastructure to quickly deploy applications and workloads to support your interest.

How much can you save? As various online calculators show, the cost savings can be impressive. Of course, in each case - depending on the provider and the needs of the customer - different results are obtained.

  •  

RoyJones

Do you offer dedicated servers, perhaps as part of another project?
  •  

arthyk

Obviously, with the development of cloud technologies, companies gradually no longer need to create their own data centers. This is simply not profitable, because reliability and fault tolerance do not suffer at all, but resource costs are reduced. In addition, the trend towards 'globalization' in this area (Amazon, Microsoft and Google, in fact, own most of them) will continue to force larger companies to use ready-made solutions.
Interesting article, thx. :)
  •  

JustinC

The server has such a concept as uptime — the time of uninterrupted operation of web server from the moment of launch. Server uptime can also be measured in relative units — percentages.
How is server uptime calculated as a percentage? From 100%, we subtract the server downtime divided by the entire server operation time. What does server downtime depend on?:

server software failures (log overflow, memory leaks, hard disk overflow);
server hardware failures (If you host your server, then its repair is the client's concern by default, or our support carries out repairs for a fee. If you rent a server, we either repair it promptly, or issue a new one.);
interruptions in the operation of the data center:
failures or accidents on the cooling system;
accidents in the power supply system;
accidents on routers and communication channels, DDOS attacks;

The server never has an absolute 100% uptime. The server will need to be changed sometime. We can talk about local uptime for a given period of time and, as shown above, it is necessary to distinguish at least three different types of uninterrupted operation.
Accidents have been, are and will be everywhere and always. Some cases can be minimized, somewhere it is possible to reduce the time of their elimination, but it is absolutely impossible to get rid of them. It is possible to minimize the human factor, but it is impossible to get rid of the failure of terminal equipment. To increase reliability, we can advise you to have two power supplies on the server and power both of them, and from different sources. Connect both network cards from different switches. And the most correct thing is to place a backup server in another data center.

Google, for example, assumes that any server can fail. According to the company, they do not use raid arrays, expensive fancy servers. They use a lot of medium web  servers. Their ideology is that there should be backup servers.
There is a direct relationship between the stable operation of the data center and its Uptime Institute certification according to the Tier III standard. Data centers that do not have all three Tier III certificates have a large number of accidents.
  •