Editor’s note: We offer our long-term sponsors the opportunity to write posts and tell their story. These posts are clearly marked as written by sponsors, but we also want them to be useful and interesting to our readers. We hope you like the posts and we encourage you to support our sponsors by trying out their products.
The cloud promises many potential benefits like scalability, increased flexibility, greater efficiency and higher degrees of transparency. It is easy for those benefits to be lost or replaced with new problems if you are not careful. Done right, migrating to the cloud should preserve those benefits as well as giving you a quick return on investment. Here are five tips on how best to make the transition.
Minimize the Upfront Labor Needed to Migrate
Cost savings from using the cloud are significant and well known, however a complicated and lengthy migration can quickly erode those future savings by replacing them with up-front costs. Choose a platform that doesn’t make you change the way you are working. Try to migrate your infrastructure one to one. This is quick and easy, and you will immediately see cost savings for little up-front investment. You can then concentrate on tuning in your new cloud infrastructure for best performance; only then should you look at restructuring your infrastructure to take advantage of increased flexibility, scalability, etc.
As every manager knows, the true cost of IT is the labour associated with setting up and maintaining the hardware and software. Avoiding hundreds of manhours for a cloud migration is the single most important action you can take for making the cloud work financially for you.
Only Buy the Resources That You Actually Need
Most people when doing cost comparisons look at the prices and the amount of resources they need and then compare this to their existing costs. The reality is that almost all platforms bundle their resources together in pre-determined ratios. Anyone who has ever bought cable TV will know how bundling works and that you end up having to purchase a lot of product you just don’t need or particularly want. Need more CPU? You are going to have to buy some RAM and storage with that too!
Resource bundling is a great way for providers to sell you resources that you just aren’t going to use. This can really eat into your efficiency gains. Avoid resource bundling! Choose a platform that doesn’t artificially force users to buy resources bundled together.
Know What You Are Purchasing; Know How it Performs
Try to use a provider that offers transparent pricing. The combination of transparent pricing and flexible resource purchasing allows you to accurately estimate your costs and to keep them directly in line with your usage over time.
Not all computing resources are the same! 50Gb of storage on one platform may perform entirely differently to 50Gb of storage on another. In the same vein the actual computing you can get done with 1GHz of CPU varies enormously among providers; many won’t even tell you how much CPU you are actually getting! The solution is to test and benchmark. When you do, just make sure you are using a fair test for your usage and not some arbitrary benchmarking software that may only be loosely applicable to your day-to-day needs. A great benchmark is useless if it is testing different actions to your regular usage.
Use evaluation periods to set up part of your infrastructure that you intend to migrate, and see how it actually performs. Compare this with other platforms. Compare the computing achieved against cost – not the price for resources. A more expensive but higher performing platform could be the cheaper choice as you will purchase less resources but get the same or more computing done.
Many providers offer cheap headline prices backed up by low performance from oversubscribed infrastructure. You end up purchasing more cheap resources to get the same work done. Don’t get caught out by this trap. Try before you buy and compare cost against actual computing throughput.
Avoid Lock-In in All Forms
There are many forms of lock-in; none benefit you as a customer so avoid them. Data lock-in is common but usually quite visible prior to signing up. Ascertain with any potential provider what the exit strategy for your data is, how quickly this can be achieved and at what cost. Is the provider savvy to this concern? Do they make it easy for you to leave? If not, maybe this isn’t the right provider for you.
System lock-in is also common but less well understood. Any platform that requires you to invest a significant amount of time changing how you work to conform to their system is creating a system lock-in for you. Unless their way of working is the industry standard, should you consider moving you will face the prospect of a major re-design of your cloud infrastructure and potential disruption. Avoid system lock-in by choosing providers that participate in open standards and whose system does not deliberately force you to change the way you work in a major way.
What Physical Infrastructure Will My Cloud Be On? What is the Redundancy?
No one wants to migrate to the cloud only to find they take a step down in terms of reliability and availability. Beyond provider competence, the physical infrastructure that a cloud relies on goes a long way to determining these two factors. A provider using legacy hardware without redundancy cannot hope to provide the level of reliability and availability of a provider using new equipment with multi-level redundancy. This holds true however big and experienced they may be. No one can turn a Lada into a Mercedes! Some key questions to ask are:
- What server equipment do you use? What CPUs are deployed? (Use this as a good indicator of their age.)
- What power redundancy does the data centre offer?
- What power redundancy do you have at the rack level?
- What power redundancy do you have at the individual server level?
Ask the last three redundancy questions for networking as well.
Power and networking outages are the biggest causes of problems. Having robust UPS-backed dual-power supplies to all equipment as well as multi-redundant networking will ensure high availability and high reliability.
Look for a provider that is transparent and open about their infrastructure and ways of working.
Following the above tips will help you avoid some of the pitfalls when migrating to the cloud. In general, look for providers that are open, flexible and transparent about the issues raised above. By doing this you can make a smooth transition to the cloud, keep your options open and truly benefit from what the new world of cloud computing has to offer.
CloudSigma is a pure Infrastructure-as-a-Service provider based in Zurich, Switzerland. They offer an open platform that combines complete control of the software and networking layers with direct FTP access and transparent billing. They currently offer a 14-day free evaluation period of their platform with substantial computing and networking resources included. For more information visit www.cloudsigma.com.
Photo by Bruce Clay.
