David Linthicum is the Senior Vice President of Cloud Technology Partners.
The cloud has revolutionized the way we build IT systems within enterprises. Indeed, enterprise IT’s goal since the inception of cloud computing has been to replicate the power of cloud computing within their own data centers.
The trouble is that cloud computing systems were built net-new, which meant they could start from scratch and thus be more innovative with the use of cloud-based resources using the most modern technology and approaches available. Enterprises don’t have the same luxury. Decades of enterprise hardware and software purchases exist at diﬀerent levels of maturation, and those structures must also support mission-critical systems in operations.
However, things are changing. New technology now provides enterprises with the public cloud experience, which includes:
- Elastic use of compute resources, such as storage and compute.
- Metered resource charge-back, meaning you only pay for the resources you use.
- Auto- and self-provisioning; you can spin up and spin down resources as you need them.
- Tight integration with new approaches and technologies, such as DevOps and the Internet of Things (IoT).
- Business agility, which is perhaps the most valuable aspect of using the clouds, means that you can quickly change applications and resources with almost no impact on operations.
In this article, I’ll take you through the steps to leverage the value of public clouds on-premises. I’ll include a path to leverage cloud concepts in ways that you may not have known about, where available new technologies support the concept of software-deﬁned data centers (SDDC).
The Public Cloud Experience
The data is overwhelming around the adoption of public clouds. By 2018, IDC forecasts that public cloud spending will more than double to $127.5 billion. This forecast is broken down as follows: $82.7 billion in SaaS spending, $24.6 billion for IaaS and $20.3 billion in PaaS expenditures. What caused the shift to public clouds? There are five primary strategic drivers:
- Purchasers believe the current cost of traditional enterprise software and infrastructure (storage and compute) is disproportionate to the value that it creates.
- In these budget-conscious times, there is intense pressure to reduce the cost of acquisition and maintenance of software and hardware solutions (the on-going support and maintenance of solutions can often be four times the original capital cost).
- Organizations strive to reduce risk, and want a far more tangible relationship between software and hardware beneﬁts and costs.
- The drive for reduced risk demands a much greater predictability of the running costs of the organization’s software solutions.
- The value of solutions is no longer determined by the functionality available, but by the feelings and experiences of the users in the way that they use and interact with the solution. In fact, most organizations only use a small subset of the functions available in their software products.
Tactical advantages are easier to deﬁne:
- Cloud uses a pay-as-you-go model to access a variety of IT resources, and this allows enterprises to only consume the needed resources.
- On-demand access to resources at a low cost allows leverage of resources when and where needed, at a fraction of the cost of owning hardware and software.
- Resource elasticity meets varying demands, allowing expansion and contraction of resources as needed.
- Colocation of computation and data enables large-scale data analytics. Systems that were once out of reach are now aﬀordable.
One of the primary drivers of cloud computing is the value of business agility. Business agility is the ability to make quick changes in a business to meet changing business needs. Examples would include the ability to add a new product line, expand into new markets, or provide customer visibility into product shipments.
Public cloud computing provides business agility. The ability to provision and scale a system is built into the architecture of most public clouds. If there is a business need for a new system, it’s just a matter of provisioning the resources required from public cloud providers. This process is much quicker and easier than purchasing, conﬁguring, and hosting your own hardware and software assets.
The value of business agility really depends upon the type of business. Those in the healthcare and ﬁnance verticals obtain a great deal of value from agile platforms, such as cloud-based platforms. Verticals with relatively static business processes, such as many manufacturing organizations, may not realize as much value around the use of cloud computing.
Moving forward, public clouds are becoming much more feature rich, with additional capabilities that meet or exceed what enterprises currently run in-house. This lead to public clouds being the desired platform of choice, although enterprises may be limited as to aspects of the public clouds that they can actually use.
Building an On-premises Private Cloud
So, what are the required capabilities of on-premises private clouds that will meet your needs? A few features to consider would be:
- Converged compute, storage and networking with automation and management
- Scale out on-demand with any size x86 servers
- OpenStack standard APIs delivery for operational eﬃciency and simplicity
- Technology for smart resource provisioning and allocation
- Distributed architecture with self-healing for high availability
What’s key about the items above is that they are all features that can be found in an SDDC as well as a private cloud. In essence, you can ﬁnd an analog for public cloud technology within your own data centers, including the ability to provide elasticity, centralization, and operational eﬃciency.
Moreover, these systems can provide the ability to auto- and self-provision compute and storage resources, allowing you to provision resources at the application level. The alternative is to try to predict capacities, and attempt to align need with hardware resources. It’s a guessing game that costs big money, since there is no way to tightly align the needs of production with hardware resources. However, using a private cloud within an SDDC, you’re able to provide a platform that’s able to better respond to the application and user needs.
Finally, there is the ability to support widely distributed architectures that provide self-healing capabilities, such as working around a server or storage system that’s failing. Or, working around breaks in network services, as well as dealing with active/active data redundancy in support of business continuity.
The ﬁrst step is to understand your own requirements, and that means doing the up-front work. While your own requirements will diﬀer a bit, the following, at the very least, should be understood.
- Data properties and usage. Know where your data is, what it is, and how it’s leveraged within applications.
- Storage properties and usage. How are ﬁles stored, when, where, and by whom? What applications exist on, or are using which storage services?
- Security and governance services needed. What are the security systems in place, and how do they need to exist in the to-be architecture using a private cloud and SDDC? The same with governance services, such as service or API governance.
- Application portfolio and proﬁles. What applications should and need to be moved to the new private clouds platform, and which ones should move ﬁrst? Moreover, understand their links to data as well as how users access the applications.
- Networking and other infrastructure needs. What is and will be the loads on the network, and how will the target private cloud deal with the loads? There are other things to consider as well, including power management, monitoring and management consoles, and other aspects that exist within your premises.
Architectures and Solutions to Consider
Once we have the requirements down, we can consider a number of meta-architectures as options including: Private, public, and hybrid/multi. Public has already been explored above. Private clouds provide you with more control and sometimes better eﬃciency as well, if the right technology is selected and leveraged in the right ways. In other words, you can replicate the public cloud user experience by using private cloud solutions.
However, sometimes it makes sense to pair private clouds with public clouds, thus creating a hybrid or multi-cloud solution. These are more complex than just private or just public clouds, but do allow you to place diﬀerent workloads on diﬀerent clouds, depending upon what those workloads need to do. For instance, placing a big data system on a public cloud for cost eﬃciency due to storage needs, and having that big data system work in conjunction with systems that exist on private clouds.
Of course, there are many public cloud success stories, including Netﬂix. They focused on delivery of a video streaming service that was more cost eﬀective to run on a public cloud. Indeed, they are going so for the same reason that Groupon is running in their own data center. Netﬂix ran the numbers, and, for them, the public cloud was the most eﬀective and eﬃcient solution.
What are the Must Haves When Looking for a Private Cloud?
So, what are the must haves when looking for private clouds solutions? There are a few key concepts to consider.
First, ease of use. The private cloud technology must serve those in operations, as well as application developers, and even application users. There are many dimensions to consider, including the ways that each person views the private cloud through their own set of interfaces.
Second, cost eﬃciencies. How well does the private cloud do at minimizing costs? It’s important that we run the numbers and understand the cost expectations, in terms of what we’ll spend for the private cloud by resource, application, and user, as well as what eﬃciencies the private cloud will bring.
When doing these calculations, make sure to ﬁgure in the value of agility, or the value of providing the ability to quickly change and expand business processes. While this value is diﬃcult to determine, you should look at the strategic value of agility to the business. For example, what is the value of bringing a product to market in days, rather than months or sometimes years? It typically means millions to the bottom line.
Call to Action
So, what does all of this mean for your enterprise? The core message: The value of public cloud does not necessarily need to be delivered by a public cloud. Using SDDC and converged systems, as well as best-of-breed private cloud platforms, enterprises can provide a cost-eﬀective hybrid cloud alternative to public clouds for enterprises.
As cloud computing continues its growth in popularity, and public clouds become more powerful, the use of the public cloud resources will be contraindicated by many enterprises. For those companies that exist in this category, it’s good to know that they have powerful and eﬀective options that will open the door to the public cloud experience.
Industry Perspectives is a content channel at Data Center Knowledge highlighting thought leadership in the data center arena. See our guidelines and submission process for information on participating. View previously published Industry Perspectives in our Knowledge Library.