Whatever industry you're in, being able to boost business agility and speed time to market in a competitive environment is key. IT plays an increasingly pivotal role in helping you meet these goals, with 98 percent of IT leaders1 believing IT is critical or very important to business strategy. However, traditional IT systems were not designed with today’s flexible workload use cases in mind. Rigid infrastructures and complex legacy systems can frustrate the ambitions of your innovative developers and IT experts, ultimately hindering business growth. This means it's important to think about where your workloads should be running, and what deployment model will offer the agility you desire.
For many organizations, the solution is hybrid cloud. Potential benefits include cost optimization, greater scalability and business agility, as well as the opportunity to bring in outside expertise and knowledge from your cloud service providers.
Today we’ll outline the four most common considerations when choosing where to place your applications in a hybrid environment.
Consideration 1: Demanding Latency and Performance Requirements
While strong performance and low latency are desirable for almost all applications, Intel customers report2 that they are particularly important for:
• Solutions that must be close to an end user’s location (e.g. online retail transactions, where the end user may be in one part of the world while the data they're accessing may be in another)
• Resource-intensive transactions with high compute, memory, and I/O requirements
• Applications with tight Service Level Agreements (SLAs) around response times
If your applications and databases must respond to customer requests within a defined time frame, or they require very high availability or redundancy, consider the hybrid cloud, keeping primary data sites on premises while locating secondary data sources in the public cloud. This model would allow IT to more easily control and predict future performance and networking requirements, and also provides you the high-availability and disaster recovery required for business continuity.
When using public clouds for performance-hungry workloads, check your cloud service provider’s (CSP) policies and cost around compute, memory, data storage, and migration. Whether you choose to use private, public, or a mix of both clouds to host your high-performance and low-latency workloads, choosing the right hardware for these applications is another very important consideration. Running these demanding workloads on modern hardware is critical to ensuring you get the performance and flexibility needed to drive today’s business success and future growth plans.
Consideration 2: Security and Regulatory Concerns
A typical organization today is responsible for a large – and fast-growing – amount of sensitive data. Protecting it is imperative, both for the company’s reputation, and to ensure compliance with diverse and complex industry regulations. With companies today using an average of four different private or public clouds3 as part of their IT environment, it is important to ensure your security requirements and obligations are met consistently, and across all clouds.
As you consider which type of cloud environment will work best for each of your workloads, make sure you identify the security tools, processes, and controls you will need - for example, encrypting data both in transit and at rest. You may wish to extend existing policies and tools from an on-premises or private cloud environment to any public cloud installations as well. In this case, your IT team must be able to manage this as more workloads potentially move to and from public cloud environments in the future.
Additionally, it is critical to ensure your cloud services offer the mature tools required by your business leaders and your organization’s security policies. For example, some regulations require standards-based federation of identities and accounts, so it’s important to ensure access to applications is kept simple for users while also ensuring compliance to policy and regulations.
Understand not all tools will work in every environment, so when you transition workloads to the public cloud it’s important to ask about the technologies, protocols, and processes underlying their own security capabilities. And while understanding your CSP’s security strategy and capabilities is important, remember also that ultimate responsibility for your hybrid cloud security lies with you. With data potentially spread across a number of clouds and coming from a wide variety of sources and end points, your own IT team will have the best visibility across the disparate environment, and must ensure consistent, seamless security measures are in place throughout.
Consideration 3: Large Data Volumes and Storage
As data volumes increase, it’s important for all your applications – wherever they sit – to be able to access the data they need quickly and securely. Organizations must ensure the right data is in the right place, and that the applications or services that need the fastest access have the fastest response times from your storage. As edge compute becomes more prevalent, this issue of data storage and volume management is becoming more of a priority.
Start by evaluating your workloads to identify those that are most and least data-intensive. Those that create high volumes of data are suited for private cloud environments, where you can have greater control of your data and how it is accessed.
However, when there are business reasons to choose a public cloud solution for applications with large data volume, ensure your CSP is able to provide a service that scales with your business over time, and make sure you understand the costs associated with over- or under-provisioning within your environment. Another consideration in selecting a cloud provider are the costs involved in data mobility and migration. For workloads that require more data, or data that needs to be very portable, discuss the costs associated with your CSP.
For data-intensive applications that contain time-sensitive data (i.e., that which must be delivered in lightning-quick time to support customer-facing or mission-critical applications), be sure it is on the fastest storage, with fast network connectivity. It should ideally also be in close physical proximity to where it will be used, to help minimize latency. Consider investing in fast storage technologies like All-Flash NVMe*, which enable faster data transfer compared to traditional hard drives4, and therefore low latency. Also keep in mind that these are important considerations no matter what cloud environment you choose to host the app, whether in the public or private cloud, fast storage and networking technologies are necessary for quickly delivering large amounts of data to end-users.
Consideration 4: Multi-cloud Integration
Getting all your systems, applications, and storage resources to talk to each other can be a challenge. In hybrid cloud environments, the sheer number of integrations needed can increase complexity and costs, to avoid any unexpected costs or risks be sure to define a holistic plan that includes your entire application landscape from the outset.
Running multiple clouds that can’t talk to each other is never ideal, so establishing smooth communication across your hybrid cloud is priority one. The benefits of being flexible enough to adapt to changing demands by spinning up new hardware and services, with minimal friction or re-engineering, is critical for hybrid cloud success.
Having a single pane of glass management interface for all your workloads is key as it offers more policy control, greater automation, and the ability to log and manage information more easily with smoother single sign-on access. Luckily, as hybrid cloud solutions have matured over the years, there are many proprietary and third-party software choices in market today. Using a hybrid cloud orchestrator can decrease the cost of setting up and managing these complex environments that span multiple clouds. Tools like cloud management platforms and automation frameworks are key to making this happen.
However, what not everyone considers is that the underlying infrastructure of your cloud instance is also critical to achieving cloud portability. For example, building upon common Intel® architecture provides you more flexibility because industry leading ISVs have optimized their software for Intel hardware, such as VMware vSphere vMotion*, which allows you to move an entire running virtual machine from one physical server to another, an important feature when managing a multi-cloud operation.
Additionally, by building all your clouds on one common Intel® architecture, only minimal re-validation or re-testing is needed as you scale your workloads, giving you the peace of mind that you can easily extend existing applications to any public cloud built on Intel.
Deploying any hybrid cloud solution involves making choices about where each workload should go. The most influential factors in each decision should be the needs of the business, and what is available in the ecosystem. When it comes to technology considerations, some of the rules of thumb, as discussed above, are that workloads with significant performance and latency needs, security requirements, and integration with legacy apps are often well-suited for private cloud. Public clouds provide immense benefits for building and scaling compute intensive and variable workloads, maintaining business continuity, and providing a development environment to explore emerging workloads such as artificial intelligence.
To understand in more details which workloads will work best in which type of cloud for your organization, you can use the Intel Cloud Affinity Model for Workload Placement.
Every hybrid cloud environment is different, meaning each one will have its own combination of priorities and opportunities. By properly considering these four areas, you can ensure you’re on the road to hybrid cloud success.
To learn more, watch this video on hybrid cloud myths and misconceptions.