Colocation and the Cloud: Just Someone Else’s ComputerDevOps by Edward Moy Tags : amazon web services, azure, cloud, cloud provider, colocation, google app engine, public cloud
Every business runs on information. Yesterday’s filing cabinets and file rooms have evolved into an infrastructure which includes software, storage, networking, security, updates and patches. Today this infrastructure could exist a continent away inside modern fortresses filled with a massive array of servers provisioned with network, power, environmental and security services.
Customers can buy their own servers for their business and build their own network, install software, obtain access to bandwidth, environmental control and electrical power. This involves upfront costs and higher technical requirements to maintain 24/7 operational effectiveness.
(Photo credit: Univac Server Room via wallpapervortex)
Most businesses don’t have to do this — instead they will use Colocation and/or The Cloud.
Colocation in a data center is not unlike renting an office or apartment and the tenants residing inside are servers and equipment.
Alternatively, businesses can also “rent” everything from a Cloud provider. Customers are paying for on-demand computing within the infrastructure of a Cloud provider — and their machines may be made of software: virtual servers on providers’ machines. This idea has existed for decades but has only been implemented in the last few years.
The earliest concept of “the Cloud” is credited to Joseph Licklider’s efforts to create the Internet’s predecessor, the Arpanet, in the 1960s – but it wasn’t until 2011 that an official definition of the Cloud was issued by the National Institute of Standards and Technology (NIST):
Cloud computing is a model for enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction.)
And what can customers expect from the Cloud according to NIST?
The NIST definition lists five essential characteristics of cloud computing:
on-demand self-service, broad network access, resource pooling, rapid elasticity or expansion, and measured service.
It also lists three “service models” (software, platform and infrastructure), and four “deployment models” (private, community, public and hybrid) that together categorize ways to deliver cloud services.
Some of the world’s biggest companies are providers: Google, Amazon and Microsoft.
While each company is known by the public for their main businesses — search, shopping and office software — and they also built an infrastructure to manage massive amounts of data.
As the Internet grew and large scale computing costs became cheaper these giants repackaged what they built at scale into a shared service – “cloud computing” – which could be sold to customers as a service. The most common version of this service is the “Public Cloud”. Some of the hottest startups have been able to grow quickly and operate 24/7 thanks to the Cloud.
The “Public Cloud”, on-demand I.T. services and resources for sale to all, is relatively recent:
In August 2006, Amazon created subsidiary Amazon Web Services and introduced its Elastic Compute Cloud (EC2).
The total market for Cloud usage will grow to a 300+ billion dollar market in just a few years, according to a Gartner forecast which “projects the market size and growth of the cloud services industry at nearly three times the growth of overall IT services”. Part of this shift is also generational attitudes about the Cloud – millennials’ came of age at the same time as the Cloud.
Despite the Cloud’s success and growth you may not necessarily choose this solution.
Before the Cloud there was another solution businesses used for years: Colocation.
Remember that the “Cloud” is “just someone else’s computer” – as Jeff Atwood, the founder of Stack Overflow, reminds in his blog “Coding Horror”. Likewise, Colocation involves machines housed in data centers which are essentially the hotels where your servers and related equipment will stay — provided space, power, cooling, and network connections. It’s not unlike a real estate rental for business tenants in need of space and common resources like power.
Jeff Atwood shared his thoughts about the economics of colocation versus Cloud services
“There’s no denying that spinning servers up in the cloud offers unparalleled flexibility and redundancy. But if you do have need for dedicated computing resources over a period of years, then building your own small personal cloud, with machines you actually own, is not only one third the cost but also … kinda cool?”
Atwood’s view, after an experiment of comparing the cost of using “Mac Mini” servers for colocation with a cloud version using virtual machines, thought it might be cheaper if users built their own “personal cloud” over long periods of time. If you needed something fast and flexible for changes and scaling up or down use, then the Cloud was an attractive solution.
But if you have the scale and the resources, you could save millions by colocation. A prominent example of tech giant Dropbox, which was a customer of Amazon for years, migrating to its own infrastructure in 2016 enabled them to save about $75 million dollars.
A Wired story covered this move:
Dropbox built its own vast computer network and shifted its service onto a new breed of machines designed by its own engineers, all orchestrated by a software system built by its own programmers with a brand new programming language. Drawing on the experience of Silicon Valley veterans who erected similar technology inside Internet giants like Google and Facebook and Twitter, it has successfully moved about 90 percent…some companies get so big, it actually makes sense to build their own network with their own custom tech and, yes, abandon the cloud.
The allure of cost savings aside, not all businesses and customers are alike – knowledge, time and finances decides which is the best computing infrastructure choice. The tradeoff is between customers’ time-frames and technical know-how versus financial resources – flexibility and redundancy.
The one thing in common with both colocation and the Cloud are shared resources. Everything will run within a data center’s infrastructure. The difference is the degree of physical assets involved: physical servers for Colocation vs virtual servers in the Cloud.
Why choose Colocation?
You might have very specific and stringent I.T. requirements about how you manage your data. Your business model may involve closer oversight of physical servers that aren’t shared. Business planning projections involve long time frames with known costs. Lastly a business may have the resources and technical specialists available to deploy a colocation solution.
Customers choose colocation as part of IT plans designed to achieve greater control and certainty. Businesses researching their options will have to account for security and regulatory demands (e.g. HIPAA (Health Insurance Portability and Accountability Act of 1996) and PCI DSS (Payment Card Industry Data Security Standard)) are non-negotiable and must be complied with. Security, safety, privacy and tighter control and configuration requirements may also include provisions for disaster recovery plans and failover sites.
What are trade-offs for using Colocation?
Fixed upfront costs for equipment for computing, storage and network connections within a datacenter – you can’t start really small and then scale up and down with your actual usage. You may even have to account for the cost of shipment and/or transportation of your equipment to a datacenter and details including relevant insurance costs. And it all must be kept up to date.
Why choose the Cloud?
You might be focused on low startup costs and freeing employees to focus on your core business – and your business doesn’t need to own servers or even have them as physical.
Using the “rental” analogy, your businesses’ data is a tenant inside a giant rental building, and you’re sharing common resources with other tenants but have your own space in this building with its own basic security and amenities.
The cloud provider is the building owner and landlord, and spent its money on maintaining the infrastructure and its upkeep. Customers can come and go as they please and may be month to month tenants in effect. Customer business’ IT budgets and teams could be freed to focus on work that is more about the actual core daily business itself.
What’s a trade-off for using the Cloud?
As your business grows, and everyone wants that for their businesses, then their data and IT infrastructure needs will grow and so will costs for using Cloud solutions.
There is no way to present a cost comparison within this limited piece – this requires a direct audit and some research of potential options and their respective costs. Business owners and management must invest the time and resources to research and plan over multi-year time frames. There is no one size fits all solution and the right and responsible thing to do is to reach out to specialists and consultants who are able to help businesses choose the most effective option for their business.