So you’re tasked with planning your organization’s cloud environment. How do you identify how much cloud infrastructure is required to meet the needs of your end users, but also deliver overall cost savings for your boss?
I’ve assembled 10 Tips for Sizing the Cloud in the following cheat sheet. It will answer many of your questions and determine your needs for today as well as planning for growth in the future.
10 Tips for Sizing the Cloud Cheat Sheet:
1. Examine your needs
- How much capacity is available in the data center?
- How much of that available capacity is currently being consumed?
- When will capacity free up?
- What is the prediction for new requests?
2. Decide which cloud model is good for you:
- Do you want to manage data center resources yourself?
- Can your company’s data reside outside its walls? Decide if all or some data needs high level protection better-managed with internal resources.
- Using a public cloud, some companies need you to sign service level agreements. Make sure you are ok with these conditions.
- What is your downturn plan? Can you quickly dispose of IT capital equipment purchased during the last growth cycle?
- How long does it take to acquire the computing resources needed to start a major project?
- How much does it cost on a CPU/hour basis to run your environment?
- Are some of your computing capacity needs tied to specific goals during a specific time frame, or are they seasonally based?
3. Think short term
Do you need the cloud for added IT firepower? That works well as a short-term strategy and is much more affordable than alternatives. However, if you require IaaS for more long-term needs, outsourcing may be your best option.
4. Take a look at your applications
Traditional applications work best running on hardware – although they can be installed on a virtual server (although not replicated across more servers). Which applications are you planning to host in the cloud? Consider API’s so your applications are automatically scalable and the resources are available as required. You may also have to upgrade your software so it’s compatible with a virtualized environment – or avoid the cloud altogether.
5. Get ready to virtualize!
Uncouple your applications from their underlying hardware to free up server capacity. For a really good guide on client virtualization, read our post Client Virtualization as a Strategy to Support BYOD, which takes a very detailed look at sizing you server, storage and network needs for virtualization.
Why BYOD? Virtualizing your client, apps and servers can enable BYOD and allow your users to access central resources remotely from any device. Also, once a server is virtualized, it can be dragged and dropped from one piece of hardware to another.
You will want to monitor:
- The number of user requests
- The number of virtual machines requested
- Your allocated CPU, memory, and disk
- Your actual consumption of CPU, memory, and disk
- Your total cloud capacity
6. Gauge performance
When it comes to performance, speed is paramount. Take a look at your cloud provider’s servers and network connections and Internet routing between your environment and the servers. Hosting proximity and location-specific factors are important. If you are planning to use your cloud to provide a service, can you guarantee speed, services and uptimes? Consider caching locally or on a local appliance that can cover up any latencies.
7. Decide whether location is important to you
Where is your data going to be located? A recent regional power outage caused Amazon’s cloud to go down. Customers running services on the cloud that were only regionally backed up had few options when the region was out of power. Ensure you are compliant – does your data have to reside only in Canada or the U.S.? Decide how much weight you are going to place on the location and privacy and any regulations that affect how you handle your data.
8. Ensure data access
Can your provider get you the data you need, even in the event of a catastrophe? What if you need to make an unusually large request for a data set? Make sure you know the copy and ship times, and if these capabilities can meet your needs.
9. Work backwards
API’s and services may not make the SaaS integration process seamless. It can be a complex process without the right approach. Ask yourself: what legacy or in-house systems do your applications need to interact with? How much consumerization or refinement is going to take place today? In the future? How are you going to provision application integration on an ongoing basis? This will give you an idea of your intial and ongoing SaaS costs – you may consider hiring a cloud integration expert or getting an expert assessment for this.
10. Know your numbers
As a general rule, if you have to meet short term needs and you don’t know how much usage you will require, cloud makes financial sense.
- Capital expenditures vs. operational expenditures
- Contract rates vs. cloud rates
- The likelihood of requiring computing power now vs. on an ongoing basis
- The provisioning costs of renting cloud space vs. an in-house solution, including capital expense, expertise, real estate, electricity, HVAC and other costs
- The costs of SaaS integration at the outset vs. on an ongoing basis
These suggestions are intended as functional guidelines for sizing your cloud environment, I hope you found them helpful!
Want to learn more about cloud? Read our first Cloud Cheat Sheet: 5 building blocks, 3 service models and 4 deployment models post. Want to keep reading? Learn about HP’s pre-configured cloud offerings by checking out Sizing your Cloud? Try a pre-configured virtualization setup or Focus on IaaS services you actually want to deliver.
Do you have a cheat for sizing a cloud environment? Share it with us in the comments below!