As cloud computing evolves, its definition seems to become looser. And while public cloud, private cloud and hybrid...
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
cloud are all logical extensions of the basic cloud computing idea, other "cloudy" variants are questionable.
For example, colocation -- the predecessor to public cloud -- is often confused with cloud. In a colocation model, a user places systems and storage in a third-party data center to take advantage of networking or a better location. The main distinction between colocation vs. cloud lies in the way enterprises use servers. And while servers are virtualized in both public cloud and colocation, the two are worlds apart.
What are the unique characteristics of cloud?
With cloud, a powerful orchestration suite controls the instances' birth and death. And this capability really distinguishes cloud from other technologies. Colocation usually involves fixed-term contracts and little flexibility, while the cloud offers compute power on-demand.
Clouds are typically built on commercial off-the-shelf (COTS) x86-based systems, which minimizes costs and enables rapid technology advancements. Clouds are also designed for multiple tenants and an ever-changing spectrum of applications.
To transform IT operations, cloud's agility is key. It allows more to be done when necessary, and minimizes costs when loading is down. The cloud also provides instances, such as those for big data, which are unavailable in most colocation equipment due to their infrequent use.
What are the benefits of colocation?
Colocation isn't as passé as you might think. In fact, for stable or long-term use cases, such as Web-serving or media streaming, it might be a better choice than cloud for some organizations.
Additionally, colocation helps resolve one of the biggest challenges associated with hybrid cloud: moving data between private and public clouds over a low-bandwidth WAN. To overcome this, users can place shared storage in a colocation facility to take advantage of faster links to public cloud. If the facility has dedicated fiber, the results are even better.
Colocation, in this context, is topologically different than using server farms to buffer content -- a formerly commonplace model. However, colocation still relies on a host to house and network the gear. As this idea evolves, expect storage rental deals from colocation providers. But because they'll offer fixed, dedicated storage through long-term agreements, these will differ from public cloud storage.
As the cloud becomes the mainstream computing vehicle, expect further evolution of the colocation concept. Almost certainly, the next move will involve colocation within the large cloud service providers' operations. This might come in the form of a dedicated -- or rented -- server pool, or even garage space for a tenant's own containerized data center.
Generally, vendors distinguish between cloud and colocation services, using different language to describe each. With cloud, for instance, server shares are called instances. But there are some grey areas. For example, there may be an orchestration suite handling colocation gear, but the intent is usually to migrate users onto a denser platform rather than an agile environment -- and contracts will reflect this.
How does virtualization differ from cloud?
Like colocation, "mainframes as clouds," such as those from IBM, are sometimes mistaken for cloud. Although there are similarities, these systems lack cloud's nimble orchestration. And the mainframe facility has a small scaling boundary compared to acres of COTS. IBM has, however, solved the multi-tenant problem with these mainframes.
There's especially a lot of confusion in the private cloud market. Many users assume that server virtualization, and tasking each virtual server with scripts, creates a cloud. But that's an inefficient use of virtual systems, and provides little of the necessary agility for modern business processes.
Much of the confusion around cloud is a mindset issue. IT has to migrate from a control-centric view to one where it sells services to users. Failing to do this will spark a rapid expansion of shadow IT, where departmental leaders figure they can get a better deal and faster responses through public cloud.
About the author:
Jim O'Reilly was Vice President of Engineering at Germane Systems, where he created ruggedized servers and storage for the US submarine fleet. He has also held senior management positions at SGI/Rackable and Verari; was CEO at startups Scalant and CDS; headed operations at PC Brand and Metalithic; and led major divisions of Memorex-Telex and NCR, where his team developed the first SCSI ASIC, now in the Smithsonian. Jim is currently a consultant focused on storage and cloud computing.
Keeping colocation costs in check
How cloud and SDN work together
Hybrid cloud hype creates confusion
Colocation services vs. cloud hosting -- which is right for you?