Deploying resources on the public cloud is incredibly easy – so easy, in fact, that even business managers can do it. But deploying resources and managing them are very different things, and most organizations are quickly discovering that as their data environments scale, so do the challenges.
Most of the issues that arise in the public cloud can be summed up under the mantle of shadow IT – the practice by which users create, and often abandon, resources without IT’s authorization or even knowledge. This can result in lost or uncoordinated data, cost overruns, security risks and a wealth of other problems. (To learn about different types of cloud services, see Public, Private and Hybrid Clouds: What's the Difference?)
But even when everything is on the up and up, the enterprise can still run into trouble merely by virtue of the fact that cloud resources are not consumed, managed or utilized the same way as local data center resources. Here then, are the top three challenges that tend to prevent cloud infrastructure from achieving its maximum value:
Tech moves fast! Stay ahead of the curve with Techopedia!
Join nearly 200,000 subscribers who receive actionable tech insights from Techopedia.
Compliance
According to Dereje Yimam and Eduardo B. Fernandez, technology researchers at Florida Atlantic University, maintaining compliance in the cloud is problematic for a number of reasons. For one thing, there is a distinct lack of common cloud reference architectures. This does not completely subvert compliance efforts, but it makes them a lot harder than they should be. With such a wide variety of architectural styles across multiple cloud providers, the enterprise is unable to maintain compliance across distributed workloads, and it makes it difficult to assess individual providers’ strengths and weaknesses before or even after data has been migrated.
Compliance can also be hampered by an inability to maintain full access and control over cloud-based environments. Most organizations that are subject to strict compliance rules will undoubtedly spell their requirements out in the service-level agreement, but without direct access to underlying infrastructure, enforcement of these requirements is a matter of trust, and violations are often detected only after data has been breached. (For more on compliance, see Beyond Governance and Compliance: Why IT Security Risk Is What Matters.)
The enterprise should also be aware that the public cloud faces unique security threats that don’t exist, or at least are greatly diminished, in local infrastructure. Most cloud workloads are hosted on highly partitioned, but nonetheless shared, hardware, so one user’s problem can impact another. And since cloud resources are often provisioned by people who simply want to get their work done, security is not always a high priority. However, one up-and-coming option - autonomic virtual monitoring - can help mitigate this risk.
Costs
It may seem strange to list this as a challenge, given that the cloud generally supports data loads at a fraction the cost of a traditional data center, but as experience grows so does the realization that the sub-penny per GB come-on offer is rarely the whole story.
In many cases, the cloud’s rapid and easy scalability is the primary cost driver. When coupled with its self-service provisioning options, hosted environments can quickly scale up and out to extreme levels, ultimately pushing operational costs beyond the capital expenses of owned and operated data facilities. This trend is most often observed in technology startups, which launch on full cloud infrastructure but eventually start building their own IT as their business grows.
[Free Study] Outage Impact Study
A global analysis of IT downtime and its impact on businesses
Download NowEnterprise executives should also realize that even though resources are cheaper in the cloud, management costs are not. No matter where an app is hosted, it still requires a technician to monitor and maintain it, which means labor costs tend to scale as cloud deployments become more prevalent. This is one of the reasons why many enterprise workloads are being handed over to managed service providers, which provide not just the infrastructure to support applications and data, but the people to oversee them. Of course, this level of service also comes at higher price points than basic cloud.
At the same time, most cost comparisons between cloud and in-house infrastructure often fail to take into consideration items like connectivity, customization, backup and recovery and a range of other factors. In most cases, the cloud still provides a lower-cost option, but it is not nearly as dramatic as the initial sales pitch suggests and, as mentioned above, these costs can quickly scale up. Public cloud management software can help streamline operations and ensure more successful, less costly, cloud implementation.
Performance
Performance in the cloud is difficult to measure because the metrics can vary widely across CPU, memory, networking and other elements. Most enterprises are challenged enough just keeping track of their own diverse infrastructure, let alone resources that may be distributed across a number of third-party systems and providers.
Compounding the problem is a lack of visibility into cloud infrastructure, which makes it difficult to assess performance characteristics of various workloads as well as the resource consumption patterns of the hosted environment. Without this, the enterprise has no way of knowing if it is getting optimal support from the resources it is paying for, nor any clear way of improving its configurations or processes to adjust to changing business requirements. Ultimately, this lack of visibility in cloud infrastructure forces the enterprise to gauge performance on the application layer, which generally does not reveal problems until the user is aware of them too.
So what is to be done about these challenges? Increasingly,the enterprise is turning toward automation to give the cloud environment a high degree of autonomy when it comes to building and maintaining the data ecosystem. As workloads become more complex and in need of faster and more dynamic support, operations will rely on too many touch points for even an army of IT administrators to handle. As today’s automated platforms evolve through artificial intelligence and machine learning, the enterprise will find that their clouds will become increasingly efficient and effective simply by operating as needed.
It’s been a tenant of technological advancement that for every challenge there is a solution. These days, the enterprise often has a plethora of solutions to choose from, which itself can be a challenge when it comes to consistently deploying the right one. But with the broad federation of cloud infrastructure and the increased prevalence of automated, abstract architectures, most organizations will find that wrong turns in the cloud can be quickly corrected while successful solutions can be expanded and improved with far fewer complications than traditional data architectures.
Not sure which cloud services are right for you? Cloud Cost Compare will profile your app workload and decide on the best cloud and template.
