So what are the advantages of cloud computing? We're going to cover what AWS calls the 6 advantages of cloud computing versus maintaining your servers on-premises. These 6 advantages are the source of many exam questions and I'm going to cover them in the following order. Benefits 1, 2 and 5, are financial benefits of using the cloud. These save your organization money. We'll cover these as a group. The remaining benefits 3, 4 and 6 are operational benefits. They improve how your organization serves customers and we'll cover them as a second group. The first two financial benefits are, 1, trade capital expense for variable expense. And 2, benefit from massive economies of scale. How does the cloud allow you to trade capital expense for variable expense? The left hand side illustrates why on-premises servers are a capital expense, abbreviated CAPEX. A CAPEX is a large upfront cost for an asset with a useful life of longer than one year. For example, buying an office building, buying a corporate shuttle to drive employees back and forth, or buying this picture of VRTX server in one go. Those are all CAPEX. Fully loaded, this VRTX server costs about $55,000. According to consultant Tom Noel's published advice on the VRTX manufacturer's website, your organization needs to buy a new one of these every 3 to 4 years to maintain their on-premises servers. The right hand side illustrates why cloud servers are a variable or operating expense, abbreviated OPEX. And OPEX is a small cost to use an asset for less than one year. For example, paychecks to employees who work in the office building that go out every two weeks or monthly, that's OPEX. The gasoline that goes into those corporate shuttles you bought is OPEX. And the pictured AWS server instances which are priced per hour, that's also OPEX. For example, this t3.medium instance costs 4 cents per hour to run on demand. Over three years of constantly running your t3.medium instance, it costs less than $800. Variable expenses should not only be cheaper, but they should also vary: It's in the name. Meaning that you can scale those costs in or out or shut them off completely. You can reduce costs by automatically stopping or terminating servers on AWS when you don't need them. You can also serve additional customers by launching additional servers, but only the minimum amount of additional servers that you actually need. You don't have to purchase a whole another VRTX for another $55,000. AWS calls this benefit "pay as you go" or "pay for what you use". From a purely economic perspective, you would think that CAPEX and OPEX costs over and assets useful life should be roughly equal. So why are AWS' OPEX costs so low? The answer has to do with another economic concept and our second benefit of using the cloud: benefit from massive economies of scale. An economy of scale occurs when an organization produces so many widgets that the cost of producing each individual widget falls. In our case AWS provides the servers that powers so many massive apps like Netflix and Airbnb. So the unit cost for each additional server is much lower than if you purchased your own VRTX servers. Our 3rd and final benefit of the cloud is to stop spending money running and maintaining data centers. On the left, you'll see that on-premises data center costs include racking, stacking, powering, and securing data centers. On the right, you see that there are no data center costs because cloud computing means having AWS manage your data centers for you. I want to illustrate the difference with just one aspect of running a data center which is physical security. Data centers are guarded by armed guards, often ex-police. And whenever I walk into one of them, these guards have to verify who I say I am, the reason I say I have to be there. And then they have to escort me around at all times. Knowing that, this is a story of when I used to work at the 10th largest law firm in the world. One night it's 2 AM. I'm in the conference room going over a case, and I'm hungry. This law firm puts out snacks in the conference rooms during the day. So I assume there are some snacks in some cabinet nearby. I open cabinet and there's a server with a webcam on top of it. So I close it and go back to work. The next day, the head of HR walks into my office and tells me that what I did was very, very serious. And I just start laughing and she asked, "what's so funny?" And I tell her, "well if data center security is so serious to you, then you want to avoid a service interruption or breach of client information, well here's $20. Get in your car, drive to Home Depot, buy a lock, install it in that cabinet and report back to me when you're done." And she walked out, I never heard or saw from her again. And realized that service interruptions and security breaches in data centers are not unheard of. In March 2021 a fire at this French data center disrupted millions of websites, knocking out French government agencies, banks and e-commerce websites. The point is that it takes a lot of resources and forethought to protect data centers from natural disasters, accidents and malicious actors. The type of resources and work that might not be worthwhile for your businesses to expend, if there are core business is not data center management. Later in this lesson, I'll describe how the AWS global infrastructure is highly redundant, which means that it can continue to perform even if it loses a data center like that situation in France. For now, just know that you can save money if you use AWS rather than trying to manage data centers yourself.
