概要
As enterprise-scale organizations continue their digital transformation — an evolution that began prior to the COVID-19 pandemic but was undeniably accelerated by it — perspectives on the storage, analysis, and leveraging of data have similarly changed.
The cloud plays a major role in this development: Research from Gartner has projected that by 2022, 75% of all databases in the world will be deployed or migrated to a cloud platform. And a steadily growing number of organizations are determining that they cannot put all of their data and application eggs in a single cloud basket, which is driving the adoption of multi- and hybrid cloud systems.
Hybrid cloud architecture can be especially ideal for enterprises' data needs by delivering the best of both worlds across public cloud and on-premises environments. But proper construction of a hybrid environment is critical, otherwise you may unintentionally make data governance and migration more difficult than ever before.
What is a hybrid cloud architecture?
Teradata defines hybrid cloud as the pairing of one of the major public cloud platforms — Microsoft Azure, Amazon Web Services, Google Cloud, etc. — with dedicated on-premises data infrastructure. Some view it as the combination of a public cloud provider's resources and a private cloud environment overseen solely by the organization using it, the latter of which can become rather costly to maintain. While the definition may vary slightly, depending on who you ask, hybrid cloud is becoming popular among organizations that want to take advantage of cloud resources without shifting the entirety of their workloads onto them. According to IDC, 90% of all enterprises will be using some mixture of public cloud and on-premises infrastructure by 2022.
In some cases, you may see the terms "hybrid cloud" and “multi-cloud" used in a way that seems interchangeable on the surface. Sometimes, you may even hear the term "intercloud," which further confuses matters as you're aiming to understand the nature of these solutions. So, it's important for us to make clear distinctions here, point by point:
- A multi-cloud setup is one that uses multiple clouds, full stop. If using the term on its own, it will often be assumed that no cloud involved in a given multi-cloud setup is connected to any on-premises data resources whatsoever.
- A hybrid cloud solution can use resources from one public cloud service or several; what makes it hybrid is the connection to on-premises. Teradata, in most hybrid deployments, utilizes an architectural model in which the database management system (DBMS) – Teradata Vantage – and the data and applications connecting to it may be in the cloud, on-premises, or both.
- Lastly, intercloud refers to the use of services from multiple cloud service providers (CSPs) and the resultant connectivity between different clouds that allows data and workloads to be moved from one to another.
Coupling one of the major cloud infrastructure platforms with your on-premises resources ensures that you maintain the value of all you've invested in the latter while keeping your organization apace with the increasingly cloud-centric business world. As enterprises — across the globe and throughout all industries — continue to grow and the distributed workforce steadily becomes a more prevalent trend, hybrid cloud computing can help ensure that you're making the most of all data tools at your disposal. This, in turn, allows for more robust, efficient analytics that help drive effective business decisions in the best interest of your bottom line.
The benefits — and limitations — of hybrid cloud architecture
Forming and implementing the best cloud strategy for your organization is no simple feat. The decision you make could well define the operational future of your business for years to come. In considering the various options, it's critical to look from an objective perspective at all pertinent attributes of each — both positive and negative.
Hybrid cloud benefits
Adopting hybrid cloud architecture is the "having your cake and eating it too" approach: You preserve the on-premises tools and infrastructure that you've spent no small amount of organizational capital to acquire and maintain, and can use these resources (data storage, application customizations, networking, et. al.) as a repository for business-critical information. At the same time, the large-scale storage and fast pace of innovation that public clouds offer is right at your fingertips. This helps minimize the possibility of any data becoming siloed from the rest of your organization and maximizes data accessibility.
There may also be immediate and positive cost-savings effects. Spending money to enhance your network's capacity or implement an entirely new and separate one just to handle the demands posed by certain apps or services won't be necessary, as you can leverage the cloud to support these tools. Offloading data from your premises to the cloud can reduce the burden on your company's on-premises infrastructure, allowing you to save or repurpose it for specific but critical purposes, such as supporting legacy tools that cannot run without dedicated servers.
Hybrid cloud's greatest benefit is arguably the level of flexibility and control it offers. You determine the location of workloads based on organizational needs that may change frequently. You give developers the opportunity to leverage the most cutting-edge tools in the cloud. Also, when paired with the right data platform, hybrid cloud architecture grants unparalleled flexibility and cost-efficiency to your analytics workload, a benefit that can't always be realized with cloud-only data warehouses.
Potential drawbacks of hybrid cloud
Implementing a hybrid cloud strategy will be no easy task if you don't establish firm control of and visibility into your data. Latency is a common hurdle to overcome, especially if there is significant physical distance to span, or if applications are relatively “chatty” with a lot of back-and-forth acknowledgement communication.
Along similar lines, oversight and maintenance of on-premises resources can be somewhat time-consuming no matter what. When these tasks must be done while also ensuring there is no redundant overlap or incompatibility with cloud resources, things may slow down even more.
When is hybrid cloud the best option?
A hybrid cloud deployment will be beneficial for any company with on-premises data sources and infrastructure that it needs to keep in place while striking the right balance between reasonable costs and effective performance. Moving data and workloads to the cloud can massively disrupt day-to-day operations if not thoroughly planned and tested in advance. The key is to enjoy those advantages while still making the leap into cloud computing and all of its incredibly exciting — and seemingly ever-expanding — possibilities. Any organization that prizes flexibility can benefit significantly from a combination of public cloud services and on-premises data tools.
Hybrid cloud infrastructure is ideal when an organization needs to upgrade certain applications and resources in a short span of time. Some of the most popular and valuable apps, like Salesforce, Canva, and Zendesk, run exclusively in the cloud, and if you've fallen behind on adoption for whatever reason, it's critical to quickly rectify that shortfall. Hybrid cloud architecture makes it easy to upgrade to these apps and/or integrate them with existing apps without the need for new on-premises infrastructure. Additionally, the cloud architecture can boost the performance of those pre-existing applications if they were formerly running on outdated infrastructure.
Scalability is another factor that can accentuate the value of hybrid cloud architecture: When a massive spike in traffic occurs related to a new project or acquisition, or a more predictable seasonable uptick, you can quickly scale up in the cloud in a manner that isn't as feasible on-premises. Downscaling can be accomplished in just as prompt a fashion when such extra resources are no longer required.
In a nutshell, companies that want to find the best solution for workloads that each have different needs and specifications would do well to consider a hybrid cloud model.
Creating hybrid architecture: 3 key considerations
An effective hybrid cloud deployment will involve the ideal combination of CSP, on-premises infrastructure, data platform, and security. To help you determine the right building blocks of your hybrid cloud environment, consider the following factors:
- Applications and data gravity: First, determine how much data is currently cloud-hosted and how much is on-premises. If there are major clusters of data gravity on-prem, can they and their corresponding applications be moved to the cloud with minimal difficulty? Keeping data and applications together is preferred when it comes to performance.
- Bandwidth and egress charges: It's ideal to minimize outbound cloud traffic as it may be costly due to CSPs' data egress charges. As such, it is important to design core analytics operations with these financial considerations in mind. The ability for your analytic platform to make smart cost-based choices on where workloads are run can make a world of difference when it comes to the monthly bill.
- Security: Your cloud architecture will only be as secure as you've made your organization overall, and security also affects how certain data workloads are implemented. Always use strong network security, enact a uniform data governance policy for both on-premises and cloud-hosted data, and use a data analytics platform that's built with security in mind to oversee it.
Teradata Vantage is the connected, multi-cloud platform for enterprise analytics. It has the security, elasticity, and flexibility you need to grow from start to scale. With Vantage as a foundational part of your deployment, hybrid cloud architecture can enable your organization to reach new heights of efficiency and innovation.
Learn more about hybrid cloud