With the number of companies utilizing the cloud, one would think that trust in the cloud is at record levels. In fact, trust in the cloud is stagnating even as investment in it keeps multiplying. A recently released Economist Intelligence Unit (EIU) study, commissioned by Google, of 452 senior executives worldwide found that while 38% of Enterprise IT is invested in cloud technology, only 16% of respondents expressed a high degree of trust in the cloud (good write up on the study can be found here, full study can be found here).
The reasons for this can likely be found in the seemingly daily news stories about high-profile cloud outages. There was the Google Compute Engine outage in April, the Salesforce NA14 outage that dominated the headlines in May, and more recently BT in the UK. Let’s not forget the several high-profile DDoS attacks bringing down Docker and the Library of Congress. However, the interesting thing about the study is that the companies who DO put trust in the cloud tended to have higher profits, market share, and revenue growth. To quote the study: “The relationship between trust in cloud technology and positive business outcomes at high-trust organizations appears to be linked to their willingness to foster business transformations that leverage what the cloud offers. Put simply, higher cloud trust appears to facilitate behavioral and process change within an organization.”
So how do companies learn to stop worrying and love the cloud? Certainly, familiarity and use come into play. The cloud is still relatively new in many organizations. The EIU study suggests that incremental success builds trust in a flywheel type paradigm. Our work with customers suggests two approaches to demystifying the cloud and thereby overcoming the cloud trust “fail:”
- Unlock the “black box” of cloud performance; and
- Architect for cloud resiliency.
Let’s parse these out a bit.
While there are several good services for tracking the performance of public cloud compute, content delivery networks (CDNs), cloud storage and managed DNS service providers (we tend to use CloudHarmony), these services only go so far. For example, they won’t show how your users are connecting to those cloud assets. Naturally, we think Internet Performance Management (IPM) tools offer some compelling value in this department, especially in terms of getting to a more granular view of how users connect to various cloud POPs (including where those users are located in the first place!) An excellent O’Reilly eBook on the “flawed thinking” of cloud adoption, including the notion that internet performance can’t be managed, is available in our content hub: Optimizing Cloud Migration.
IPM also provides real-time monitoring and alerting of internet performance issues impacting your users, from the cloud service provider availability to the middle-mile connections between ISPs. Another use case most enterprises face is Hybrid adoption where the colocation center will stay due to legacy non-cloud ready workloads exist. IPM provides organizations the ability to manage, load balance, failover, and optimize these areas including analysis planning prior to a deployment. So the obvious next question is, how should organizations think about architecting for cloud resiliency (i.e. #2)?
The EIU/Google study also points out the human nature of linking distrust with the ability to control. At least when servers or routers in your own datacenter go down, you have the control to fix the issue (even if time to resolution is the same or even longer than that of CSPs). When cloud resources go dark, IT pros are relegated to hitting refresh on their cloud provider’s status page. Increasingly, companies are considering multi-cloud and/or multi-CDN strategies that allow them to steer traffic from one provider / POP to another (generally using DNS as the control plane). More advanced implementations introduce analytics-based dynamic steering to automatically move traffic based on real-time conditions.
Ultimately, IPM can’t address the availability of cloud resources, but it might help companies overcome their cloud trust issues to begin to reap the same rewards of those who’ve jumped in.
Gary is VP of Global Sales Engineering and Customer Success at Dyn, and Internet Performance Management company. Previously, Gary served as the Area VP for the Eastern Regional Sales Engineering at CenturyLink supporting Cloud & Hybrid IT. He also ran business operations for the company’s Hybrid Cloud business.