A homogeneous cloud is one wherever the whole software system stack, from the hypervisor or remote cloud supplier, through varied intermediate management layers, all the thanks to the end-user portal, is provided by one provider.
A heterogeneous cloud, on the opposite hand, integrates parts by many alternative vendors, either at totally different|completely different} levels by a management tool from one trafficker driving a hypervisor from another or perhaps at constant level multiple different hypervisors, all driven by constant management tool.
Selecting between homogeneous cloud and heterogeneous cloud
The argument for homogeneous environments is that as a result of everything comes pre-integrated they’re easier to line up, and if one thing goes wrong there’s just one accountable party – “one neck to wring” because the speech has it. On the opposite hand, by giving such a lot power to at least one vendor, users place themselves at the mercy of that vendor’s industrial and technical strategy. In farming, this is often called a monoculture, that is, solely one crop is grown up. Superficially, this is often a lovely plan, as farmers will specialize and cash in of economies of scale. the matter is that any illness, frost, drought, or alternative event that affects that crop will wipe out the whole harvest. To mitigate that risk, farmers try and so many differing types of crop, as insurance against losing the whole harvest.
The same arguments apply in IT. The advantage of associate IT monoculture is that everybody will specialise in that one vendor’s tools and it’s straightforward for admins to hide for every alternative. The downsides area unit slightly different: on the technical aspect, options are on the market once and if the seller chooses to develop them. the important pain usually comes on the business aspect, as a result of once users area unit secured into a single-vendor strategy, they need no recourse if that seller set to alter its rating structure in an exceedingly means that causes prices to extend.
Heterogeneous architectures commit to bypass this lock-in impact by introducing parts from many various vendors and allocating their use per a typical set of methods. At some purpose, however, one management part can get to be introduced. Defenders of homogenous approaches can counter charges of lock-in by commenting that this convergence on one management layer simply moves the lock-in additional up the stack, however still leaves users at the mercy of the supplier of that one part.
The false equivalence between platform lock-in and supposed management lock-in could be a neat rhetorical trick, however, doesn’t every inhibition. Management vendors got to carry on with the event pace of the managed platforms or risk falling behind the competition from different heterogeneous management vendors. Any try at predatory business practices is going to be nipped in the bud for a constant reason.
Meanwhile, users will migrate off from management suites a lot of simply than they will modification hypervisors or cloud suppliers. The explanation is that a platform modification is sort of certain to cause disruption unless it’s terribly rigorously managed, whereas substitution a management platform, even at terribly short warning, will definitely be painful for IT and cause delays in delivery of recent requests, however, won’t have an effect on workloads that are already running on the underlying platforms.