Enterprise reliance on cloud computing is unabatedly rising across verticals. With its pricing and scalability advantages, the cloud has become the environment of choice for cognitive computing instances, Internet of Things use cases, and application development and deployment.
It’s precisely because of these benefits that organizations are encountering similar, recurring issues in the cloud that frequently span hybrid and multi-cloud undertakings. If not properly addressed they could potentially negate the boons of this progressive architecture; if so, they can increase them almost exponentially. Common issues for scaling business throughout the cloud pertain to:
- Operations: The
cloud’s elasticity benefits are best realized when empowering
organizations to scale out and scale in. Approaches supporting both of
these capabilities, as well as what DataStax Chief
Strategy Officer Sam Ramji termed a “consistent operational stance” and
uniform application deployment pattern, are critical to maximizing
operational success in the cloud.
- Data Access: The
cloud’s portability and penchant for burst computing is ideal for
computationally intensive tasks. Nevertheless, these boons are
relatively useless without a credible means of dynamically shifting the
underlying data required for them throughout hybrid and multi-cloud
settings, too.
- Data Governance: The
converse of the gains attributed to accessing cloud data from anywhere
at anytime is facilitating the requisite permissions—and organizational
oversight of data governance mechanisms—to do so in a sustainable,
orderly manner.
By leveraging an artful combination of cloud native technologies and modern data store options, however, organizations can surmount these difficulties to regularly make their cloud deployments pay off, significantly reduce the risk of doing so, and “give you a compute and data environment that you can run anywhere,” Ramji remarked. “It doesn’t matter what your national regulation is, you should be able to get the data that you need into the hands of your users without having to worry about it.”
Cloud Native
Cloud native technologies can substantially reduce the practical challenges of cloud deployments—particularly when regularly moving data and computations between on-premises and any number of cloud settings. Cloud native approaches involving containers (which frequently support microservices) and container orchestration platforms like Kubernetes streamline the requirements for how applications are run in the cloud so organizations have “all three things: scale out [and] scale in, one developer experience, and one operational stance,” Ramji commented. Container orchestration platforms are critical to facilitating the following gains for these three capabilities:
- Scalability: In
addition to supporting what’s becoming the application infrastructure
of choice (microservices), containers are renowned for their ability to
quickly spin up applications on demand at scale—such as when
accommodating surging traffic for an influx of Netflix users during the
airing of popular television programs, for example. More importantly,
they also enable users to abdicate those resources when not needed.
“There’s plenty of ways to do scale out and provision peaks, but the
harder thing to do is to scale in so you can make the whole thing work,”
Ramji observed.
- Deployment Pattern Consistency: It’s
important for organizations to uniformly run apps regardless of their
location: whether on-premise or in the cloud. Cloud native platforms
innately provide this capacity so there’s a “consistent way for
applications to be architected, so if you’ve got a killer component you
can actually move it from one environment to another, wherever the
application runs,” Ramji explained.
- Common Operational Stance: Operational
considerations are particularly influential for the overall user
experience for end users of cloud applications. Cloud native methods
enable DevOps teams to facilitate cloud based service as though there
was just a single operating environment—when in reality, it might
consist of numerous cloud providers and types.
Data Access
Perhaps the principal shortcoming of cloud native approaches is they don’t inherently address the data issues required to avail organizations of the scalable, portable compute power they deliver. Part of the reasons containers are so lightweight and easy to run in different environments is many of them are intrinsically stateless—meaning they don’t preserve the data necessary for deploying applications. As Ramji mentioned, “Apps are useless without state.” To that end, orchestration platforms like Kubernetes are substantially enhanced by horizontally scalable, distributed data stores originally popularized during the big data movement and the emergence of NoSQL options.
DataStax recently partnered with Kubernetes to provision a Cassandra operator so that “when you write an app and you use Kubernetes, Cassandra just comes along for the ride,” revealed Patrick McFadin, DataStax VP of Developer Relations. “That’s the database you’re using by default.” NoSQL options that scale horizontally without sharding are particularly effective when used with cloud native methods, because developers don’t have to redress the code for the apps to scale (which is frequently necessary with sharding approaches). The tandem enables organizations to couple their application data “alongside the compute so you’ve got compute and data together, like peanut butter and chocolate,” Ramji reasoned.
Governance and Compliance
The aforementioned pair supports a number of impressive use cases for scaling resources as needed to meet customer demands for peak load processing across industries. However, it also helps rectify another major cloud computing complication: data governance and regulatory compliance. Regulations vary according to country and industry about where data are stored; the dynamic accessibility of the cloud may prove detrimental in this respect because it’s so easy to shift these resources and incur non-compliance penalties.
Consequently, data governance councils are tasked with determining what data should be used by which users, how, and what the quality of service will be. Ramji noted that horizontally scalable NoSQL stores can assuage these concerns by enabling IT and data stewards to “control that access” through the foregoing data store platforms. This capability is invaluable for governance and regulatory compliance, as these solutions provide a means of implementing defined business rules for reducing risk associated with data access. Such functionality immensely behooves multinational distributed operations, which is why “we see that at Federal Express and at JP Morgan Chase” Ramji divulged.
Data and Compute
Cloud native technologies and horizontally distributed data stores naturally complement one another to smooth the complexities inherent in cloud computing at enterprise scale.
This combination is able to provide a consistent, sustainable operational experience, portable data access, and the underlying data governance requisites necessary to benefit from this architecture without its shortfalls. Moreover, it enables organizations to position their data and compute side by side, which is critical for optimizing performance, user experience, and the cloud’s elasticity benefits. “One of the things we learned from the cloud is elasticity wants to work both ways,” Ramji said. “You want to stretch resources randomly, then you want to let it go.”
About the Author
Jelani Harper is an editorial consultant servicing the information technology market. He specializes in data-driven applications focused on semantic technologies, data governance and analytics.
Sign up for the free insideBIGDATA newsletter.
Speak Your Mind