Why the Enterprise Must Keep Its Head Out of the Clouds
Data Center Maintenance
To maximize uptime, many enterprises are shifting to a multi-cloud strategy to diversify their cloud reliance and, in some cases, duplicate mission critical assets across multiple providers.
The cloud has transformed the ways businesses engage with technology. Rapid scalability through elastic provisioning, increased productivity as employees collaborate from anywhere, and reductions in the IT management burden and costs are among the features compelling a mad dash to the cloud.
Are enterprises now putting too much stock in the cloud, however? There are signs that other modalities are becoming necessary to sustain the digital transformation and enable emerging technologies. As promising and powerful as the cloud may be, it behooves IT leaders to think critically about their strategies to avoid butting up against the cloud’s inevitable limitations.
Limitation #1: Access
Many enterprises turn to cloud services providers to deliver a level of reliability that can be difficult to equal in-house. Various CSPs with their geographically disparate, mirrored data centers with automatic failover and multiple internet redundancies can ensure data and applications are available in almost any circumstance. But reliance on a single provider does have its drawbacks.
This fact has been underscored by major, headline-grabbing outages over the years. For example, the Amazon Web Services interruption in 2017 affected much of the internet and was followed by a similarly disastrous U.S. outage only a year later. Microsoft Azure went down for a half day in Europe last summer and then had its longest downtime in seven years two months later. And the list goes on.
To maximize uptime, many enterprises are shifting to a multi-cloud strategy to diversify their cloud reliance and, in some cases, duplicate mission critical assets across multiple providers. This helps ensure continued access should one CSP go down, although the approach also raises costs.
An alternative is for organizations to serve as their own secondary or tertiary provider, leveraging on-premises capabilities to sustain mission critical applications in the case of a cloud outage. Maintaining an on-premises data center offers the added benefit of avoiding internet connectivity as another barrier between the enterprise and its data and services. Network bottlenecks can slow data-intensive operations, and organizations without access to backup internet providers can easily find themselves cut off from external cloud resources. These factors are among the reasons for a surge in hybrid public/private cloud strategies.
Limitation #2: Connectivity
Additional cloud downsides include latency and network traffic issues, and they are taking on greater importance due to certain emerging technologies. Many augmented reality solutions, for instance, must deliver a user experience within the timeframes of human perception, which is not possible if traffic must travel great distances from user to cloud data center and back. From a sheer bandwidth perspective, medical imaging and other fields reliant on extremely large files may not be well served by a centralized, off-site data repository.
IT organizations are beginning to pay greater attention to where data is processed and stored. For autonomous vehicles, smart cities, and other applications, more than a few nanoseconds of wait time to complete a given function will be unacceptable.
This is steadily pushing storage and compute toward the end user. More disparate locations, which are together becoming known as the edge, will complement centralized cloud and on-premises resources. Experts predict a world of much smarter IoT devices, a fog computing layer for data triage and decision-making, and local and regional processing hubs splitting workloads with the cloud.
Limitation #3: Cost
Beyond the location of data and services, another key issue with the cloud can be cost. Although most subscriptions are designed to look inexpensive at first blush, enterprises must predict their long-term needs and the cost implications.
This can make data storage, among other cloud services, unexpectedly expensive over time. Moreover, vendor lock-in can be difficult to avoid. The challenge and risk associated with moving petabytes of storage or switching mission critical services to a new provider is a strong deterrent to change, putting many organizations at the will of their current CSP partners even as prices rise beyond their expectations.
Without well attuned management of cloud provisioning, sprawl becomes a significant concern as well. It’s no wonder that 64% of enterprises consider cloud optimization for cost-savings their top priority for 2019. As the cloud matures, enterprises must regularly revisit their cloud services and their return on investment to make adjustments accordingly.
Limitation #4: Capacity
Closely tied to cost is capacity. Most companies will plan for their annual cloud needs, but with the current explosion in data volumes, it is all too easy to blow through the budget. In the IoT era in which device-generated data will help expand the global datasphere to 175 zettabytes by 2025, enterprises will invest far more in cloud storage solutions than they do today, unless they find viable alternatives.
On-premises data storage is an attractive option for alleviating the escalating, subscription-based costs of massive cloud storage. It’s unlikely, however, to be sufficient. Organizations will need to pair cloud and on-premises data storage with an informed, even ruthless, data management strategy.
No longer can companies afford to keep excessive amounts of data on hand, due to the storage costs, the security implications, and compliance rules, such as the General Data Protection Regulation. What’s more, the cloud could at some point run low on storage capacity, putting more of a premium on it.
Rather than collecting all possible information and maintaining it indefinitely lest a useful application be identified, enterprises will need to proactively chart the course of each type of data, from its origin through its timely, secure destruction. Only with a dedication to narrowing data retention activity to the information that specifically enhances operations, enables better customer service, and returns actionable insights will companies be able to survive the data tsunami with sufficient budget left to invest strategically.
Enterprises are finding that the cloud-centric operations they had once envisioned and are only now deploying in full must evolve into something yet again more complex than the multi-cloud, hybrid architectures they are pursuing. The digital transformation is driving companies to a level of data intensiveness under which network and internet resources become rare commodities, downtime an intolerable occurrence, and centralized resources inadequate to meet demand.
The cloud will continue to be a critical asset for the enterprise, but to succeed, leaders will need to keep their heads and rationally evaluate supplementary, complementary, and alternative technologies.