dcsimg
April 30, 2017
Hot Topics:

An Eulogy to Private Servers

The cloud

The days of private servers are over. Yes, there are some companies still using private servers, around 5% according to Rightscale's survey, but they can be roughly divided into two groups: those that are planning to switch to the cloud, and those that will use their current setup until it breaks and they will shift to the cloud, too. There is no reason to hype cloud services anymore because they have already become the new norm. The reality is that private servers never really had a chance once our infrastructure was good enough for cloud services.

It almost seems strange to see a company still using private servers. The tech industry is notoriously performance driven and seeing a company still use old and limited technology seems out of the norm. When we see a company still using private servers, we get the same feeling we get when we see someone's DVD or audio CD collection. You just don't realize that almost everyone around you has simply started using Netflix and iTunes so much that you've almost foregone physical media for your entertainment. Let's look at some of the reasons that the cloud has made private servers obsolete.

Economies of Scale

Economies of Scale is the most basic reason that cloud is so much better than private servers and will always be better. 'Economies of scale' is a basic economical concept that deals with the marginal cost of additional performance or production. The concept is normally explained by using the example of a factory. There are many costs that are not directly correlated with production. Let's say that you are running a factory that produces bolts. There are some fixed costs; you will have to pay rent for the factory, maintenance for the machines, payroll for the employees, even if you do not product a single bolt the whole month. The cost of producing bolts is high at this point.

However, slowly, as your production goes up, your cost per bolt will get lower and lower. The more bolts you produce, the lesser your per-bolt cost will be. This is because you already have the setup and the fixed costs are being divided into more parts. Secondly, the additional profit you get also will allow you to make your factory better. Eventually, you will be able to buy better machines that produce bolts faster and use less labor. Your cost per bolt will be even lower now. Soon, you will be able to hire more engineers to make your system more efficient. The larger the setup, the lower the additional performance costs.

Cloud servers are all about economies of scale. Think of what you have to do when you set up a private server. First, you have to buy the hardware. You will be picking up a few servers at most and you will talk to a local or a national retailer. You will be paying shelf-price for these servers. Compare this to cloud service providers. Imagine how many servers Amazon buys and how different that negotiation goes. According to conservative estimates, AWS had 2 million servers in 2014. When AWS goes to buy servers and parts, they pay very little compared to how much we pay. This isn't because we are being fleeced by the sellers; it is simply because of economies of scale. When a company sells you three servers, they have to make whatever profits they can from those three servers. When they sell to Amazon, they sell 100,000 servers at the same time, so even a profit of only a few dollars per servers is more than enough in aggregate.

That is why private servers simply will not be able to compete with cloud servers. It isn't about technological advances or pricing set by Amazon; economies of scale drive the price and it just isn't possible for a small setup to compete with something as big as AWS. Private servers aren't just the old technology in this case—they are the old technology that costs more, which is an incredibly weak position for any product. Most of the other factors that make cloud better than private servers are, in one way or the other, derived from the economies of scale as well.

Performance

When it comes to performance in context of price, private servers simply have no chance compared to cloud services. Have you noticed how quickly businesses have switched to the cloud? Usually, technological adoption would go this way: The person with technical expertise will try to convince the management that a new technology will work better. The technical person gets ignored. Finally, after months of years of insistence, the management will look at the new technology. Then, the management will decide that the costs of the new technology are too high and will continue using the legacy system. Finally, after a couple of years have passed and other managers are Okaying the switch to the new technology, the management will tell the technical expert to go through.

This whole process seems to have been skipped with the cloud, largely because of the cost. If you want to convince your management that you should switch to the, cloud it is very simple. All you have to do is find out the total cost of the private server—maintenance, equipment, everything—and then go to the management and tell them how much more they would receive from a cloud host for the same cost. The performance available in cloud hosts also beats private servers simply because the hardware used is more powerful. Cloud hosts know that performance is the key to winning clients and thus use some of the most innovative hardware as well as techniques to increase performance.

Scalability

Scalability is horrible when it comes to private servers. It is an expensive and messy process. There are two types of scalability: vertical and horizontal. Vertical is when you add more hardware to the same server; horizontal is when you have to add another server. Neither is a good way to increase the server's performance. With vertical scaling, you are severely limited by hardware capabilities of the server. Every server has an upper limit, and it is very easy to reach the upper limit.

Adding another server is a pain in itself. The company will have to ensure that the new server works perfectly with the old server. The company will also have to completely redesign the network infrastructure to properly divide the load between the two servers. We haven't even gotten to the cost yet. Server upgrades are a costly measure because server hardware is expensive. This means that management often tried to delay updates as much as possible to keep expenses down.

Scalability in cloud servers, on the other hand, is a main selling point. Scalability with products like AWS, Google Cloud Platform, and Azure is almost too easy. You simply change your package to get more processing power or hosting. The change happens instantaneously. The costs only go up a little. Upgrading a private server is an endeavor that takes months if you factor in the bureaucracy of getting the upgrade approved, finding good server hardware, and then building and testing the whole network. With cloud services it is just a few clicks.

Then comes the temporary scaling capabilities of cloud servers—a feature wholly missing and impossible in private servers. Temporary scalability means temporarily increasing your server power for a special reason. When a company is running a special promotion or gets mentioned in a big publication, it is easy for them to be overwhelmed. If you build an app or a service and it takes off on Reddit or any other similar Web site, your incoming traffic may be 30 or 40 times what it normally is. This growth will be temporary; people will want to check your service/product and only a few of them will stay. If your services are running on a private server, there is no way to manage this. You can shut down your server once it gets overwhelmed and lose out on a great opportunity to gain new customers. You can immediately purchase another private server, but it will go to waste once the temporary surge is over. With cloud hosting, you can simply purchase additional bandwidth, storage, or processing power for a short time. You will be paying a bit more only for that time and will go back to your normal billing without having to make any investments in buying new hardware.

Customization

This is perhaps the most surprising advantage of the cloud. Common sense would tell us that a private server will be more customizable than the cloud option simply because we have the server and have complete freedom to do with it what we want. This is quickly proven false when you look at the options available on a cloud service such as AWS. You have the option to pick from multiple server setups and you then can customize your own setup according to your specifications. It is also much easier to customize simply because instead of having to look for parts availability and maintenance in your region, you can click on whatever options you want.

AWS, Google Cloud Platform, Azure, and pretty much every well-developed cloud solution also have different setups for different use cases; this makes the customization even better. If your company works in media and needs to store a lot of data at a low cost, you can simply pick the configuration that uses disk drives. If your company works with a lot of data and has a lot of read/write requirements (typical in any business that processes financial data or any other data intensive process), opt for the configuration with solid state drives for instantaneous updates. There are many more configurations and you are sure to find something that fulfills your company's requirements perfectly.

Expertise

Most large corporations have a lot of technical experts for their server management. This may include, depending on the size of the firm, many networking engineers, system architects, technology manners, and more. This is not a luxury that small- to mid-sized companies can afford. A small business barely has enough of a budget to hire someone for IT support; to expect them to have a professional network engineer on the workforce is unreasonable. This means that whatever private server they opt for will not be well run or maintained. Even if they manage to make room in their budget for someone to manage and maintain the server, they will not be able to hire the best. Their system may be managed wrong or it may be designed inefficiently.

Compare this to the engineers and experts working at AWS, Google, and Microsoft on their cloud servers. Which team would you rather trust with your business data? This is another big reason that cloud servers perform so much better than private servers. The companies hire the best network and hardware engineers on the planet and give them a large budget to work with. The engineers then come up with solutions that ensure that the millions of servers work at peak performance. Most large corporations also have teams that do the same thing, but only the biggest corporations can reach the same level of efficiency.

Responsibility

When a company opts for the cloud solution, they no longer have to be responsible for the servers. This may seem odd, but it is, in reality, a huge sense of relief for the whole company. When a company is using a private server, it is 100% responsible for the server. If the server begins to slow down, they have to investigate and find out what is wrong. If the server stops working, they are responsible for replacing it. If it gets outdated, they are responsible for upgrading it.

Opting for a cloud solution means not having to worry about any of the above. They are the responsibility of the cloud provider. Your company doesn't have to worry about fixing or maintaining anything. This is important because it allows your company to focus on its own business. Cloud providers are also able to solve issues much more quickly—they have staff, redundancies, and backup servers that can solve these matters in just a few minutes.

Machine Learning

Using Cloud services allows you to innovate much better than when using a private server. This is the final nail in the coffin for private servers because this will never be possible on private servers. Machine learning is a fantastic example of the capabilities of cloud computing. Machine learning is extremely processor intensive and cannot be run on a normal server setup. Only the biggest corporations in the world can think of creating their own machine learning configuration. Customers of AWS, Azure, and Google Cloud, on the other hand, can purchase machine learning services easily with just a few clicks.

Why can't you do the same with private servers? The machine learning initiatives run on specialized hardware on these cloud hosts. When you purchase the service, the machine learning does not necessarily happen on your server, but on the specialized hardware within the same ecosystem as your server. So, unless your company is willing to spend hundreds of thousands of dollars to try machine learning, it is almost impossible.

This difference will become even more pronounced once artificial intelligence is mature enough for business use. You can already use IBM's Watson for business purposes at a cost as low as thirty dollars a month. There's absolutely nothing comparable for private servers.

Business Continuity

These days, even small businesses have clients in other locations, nationally or globally. This means that they cannot simply shut down due to a localized event. A blizzard shut down the whole city? The business still has to be active. This is a big problem with private servers with the only solution being off-site backups. This isn't a problem at all for cloud servers. Cloud servers are already in a different location and have been engineered with business continuity in mind. This is done in many ways. The first is that these data centers have the capability to properly work in extreme conditions. They are also placed in locations that are known to have a stable climate. Then, there are the redundancies; all major cloud hosts keep the same data in multiple locations to ensure that their businesses are up and running, no matter what happens.

Cloud Isn't the Future; It Is the Present

The simple reality is that cloud services are just much better at this point than having a private server. In the future, it may be unheard of to have a private server running inside the company. Each and every company that is using private servers right now should start looking at how they will move to cloud in the future. The best part is how easy the process of moving to the cloud is for most companies. Cloud servers support most of the architectures that private servers work on; this allows the servers to be shifted to the cloud easily.

We are also excited about what this means for small businesses. Small businesses, thanks to cloud computing, are now more capable than ever before. The Internet was the first technology that gave small businesses a fighting chance against big businesses. Suddenly, it didn't take a million dollar advertising budget to reach clients at a national or global level; smart companies could do it just by leveraging their social media presence and spending a little money on SEO and digital marketing.

Cloud hosts provide small businesses the opportunity to counter big businesses in the next race—performance. Previously, it was impossible for a small business to offer services that could rival services from major corporations—they simply could not afford to build the infrastructure and make the hardware investments required to deal at the same level as big corporations. Cloud services make this whole point moot and small businesses can simply scale up as they acquire more customers. It removes the upper limit in how many clients a small business can deliver services to, which is a very exciting development for the world of business.

Right now, AWS is in the clear lead when it comes to the cloud market. They started early and have managed to dominate the cloud industry. Microsoft, Google, IBM, Oracle, and many more are simply trying to catch up. This is great for us, the customers, because it means that the companies will be forced to innovate and compete just to get our business. Look at Microsoft's Cortana or Google's big data tools. They are clear attempts by Microsoft and Google to fight AWS by providing features that AWS does not. This is also an area where Google and Microsoft are using their old business to tackle the new business of cloud—both Cortana and Google's tools were made not with the express intent for the cloud but are separately developed technologies they are incorporating with their cloud offerings.

We are living in the time of transition where companies are leaving old technology and moving to the cloud. According to a Rightscale survey with more than a thousand respondents, more than 95% of enterprises are now using cloud solutions, up from 93% in 2015. A lot of new possibilities have opened up and we have access to more services than ever before. We would recommend investing in private servers only if there are some very specific demands that cannot be met by the cloud. In all other cases, cloud hosting will be a much better feature for almost every company in any industry.

We won't really be missing private servers, but we will always remember the tough times they helped us through. We wouldn't have reached where we are today if we hadn't been messing around with private servers for more than three decades. They had a good run, but now that our Internet and networking capabilities are good enough to make cloud servers a reality, good bye.

Editor's Note: This is a contributed article and the opinions expressed are those of the authors. If your opinion differs on private servers, feel free to share your thoughts below.

About the Authors

The article is written by the specialists from Adoriasoft Company, which consists of highly qualified, energetic experts driven by a mutual desire to deliver the best. The Adoriasoft team keeps on learning, growing and evolving. Ongoing self-education and corporate trainings help them be professional on the global scale. They like to contribute their expertise to specialized media sources and are always open for discussions. You can find them on social media at https://www.facebook.com/adoriasoft/ and https://www.linkedin.com/company/adoriasoft.


Tags: AWS, Amazon, Cloud, Microsoft, Azure, scalability, cloud servers, Google Cloud Platform, economies of scale, private server




Comment and Contribute

 


(Maximum characters: 1200). You have characters left.

 

 


Enterprise Development Update

Don't miss an article. Subscribe to our newsletter below.

Sitemap | Contact Us

Thanks for your registration, follow us on our social networks to keep up-to-date
Rocket Fuel