With the emergence of the cloud over a decade ago, the question remains for many data-heavy companies and their IT departments, how much information should be transferred to a cloud provider for storage and routine functions, and how much should remain on-premises in a server network? Network attached storage (NAS), storage attached systems (SAN) and content-addressable storage (CAS) are all options for storing data. Realizing the benefits of a hybrid cloud strategy, in an incremental fashion, means the IT infrastructure for a company can combine the best options to enhance its data management while leaving in-house personnel more time to focus on less mundane tasks.
What is a hybrid cloud strategy?
Connecting an on-premises (on-prem) set of servers to a cloud interface like AWS (from Amazon) or Microsoft Azure allows for data storage and routine operations that can be sent offsite. Because there is a latency issue—a delay in the transfer of data from the cloud following an instruction to do so—key operations, where speed is of the utmost essence, might require certain data files to stay on in-house servers or storage (NAS/SAN/CAS). The physical distance from the home office to the cloud provider via the internet is the cause of that latency. Policies can be programmed that would send data unused for a fixed period to be sent to the cloud for archiving. That reduces the on-prem workload for the servers as well as the storage. Companies remain flexible and can adopt a hybrid strategy as needed.
Want More Tech News? Subscribe to ComputingEdge Newsletter Today!
Since there can be a cost savings if the in-house servers are leased from providers that charge based on usage, data backup functions can be sent to the cloud. Establishing specific times when operations must run and when there is downtime can also save money since space in the cloud environment is rented and priced by usage. High ops (operations) and mission-critical data should stay on premises in many cases. Where security is an issue, as with a bank’s financial data, it might be safer to stay in-house.
Take care of the data
Network Attached Storage (NAS)—networked appliances containing storage drives—can become bogged down, “dragging the load,” as more data and additional operating routines are added. That’s when it is ideal to incorporate a cloud provider. Policies for clearing data from on-prem servers to the cloud can be established, in part, based on the last time they were accessed. Some important factors to consider when adopting a hybrid cloud strategy:
- Lessen the on-prem data burden and footprint. This can also lessen latency (or ping). For certain applications, that minute speed increase gained by reducing the on-prem server burden may be a competitive advantage. A single common repository (central file storage location) can also be configured on a local machine for a single user, but is often stored on a server that can be accessed by multiple users.
- Minimize data housekeeping. When moving data to a cloud provider, it can still be safely stored and accessible when needed. There are various integrated cloud storage systems available, including personal storage systems that may be less expensive in the long run than paying the fees associated with most cloud storage devices. The benefit, however, is that it reduces the burden on an in-house IT team.
- Lighten the data housekeeping load in-house to generate greater efficiency and cost-effectiveness. That occurs because fewer pricey on-prem storage devices are required. Cloud pricing services depend in part on how much data storage is moved there and the frequency of routines that are run. Storing archival data in the cloud can be more cost-effective than on-premise storage when considering data management issues and utilities (electricity/cooling expenses) associated with housing archival data in the on-prem server farm.
- Cloud bursting. This is a configuration set up between a private cloud and a public cloud to deal with peaks in IT demand. Economics is a key, with overflow traffic directed to a public cloud when the private cloud reaches 100 percent capacity. A company only pays for those additional cloud resources when they are needed. Basic applications can be moved to the public cloud to free up local resources for business-critical applications (according to azure.microsoft.com). When bursting to a public cloud, consider security and compliance requirements for data, platform capabilities, and whether there will be a latency (speed) issue.
- Autoscaling. A cloud feature that automatically adjusts the number of computational resources needed (adding or decreasing servers) and is a popular way for retail-focused online companies to manage cloud resources. A “warm-up” exercise can help detect where additional cloud resources might be needed when shoppers head to their phones or computers for bargains.
- On-premises consumption model. In this case, a company does not have to pay for servers it does not need based on the data burden. Consider it as a partnership between a company and the cloud service provider, reducing upfront equipment and maintenance costs for data-heavy businesses. Providers like AWS (Amazon Web Services) and Microsoft Azure can analyze in-house consumption levels to assess what tier of services may be needed. Vendors then craft a customized mix of standard and high-end servers to distribute data efficiently over this new consumption model, sending data files to the cloud where appropriate. This all happens after carefully planning a data migration strategy from the old infrastructure to avoid downtime.
Beware of pitfalls
Moving some data storage and operations to the cloud from on-prem should only come after careful analysis that includes the aspect of data latency. Some functions may have to remain stored on in-house servers. Cloud services have become popular for backing up stored data as a hedge against on-prem system crashes or security issues.
In the cloud, there’s also the serverless option that utilizes on-prem server farms when needed to perform required functions; which can result in cost savings for the cloud client. The more functions and assets utilized from the cloud, the higher the rental costs are. The cloud can mean more data agility but does not necessarily mean it’s less expensive. The routine operations that can be sent to the cloud allow in-house IT employees to spend more time working on new innovative programs as their company grows.
Cloud maturity
A decade-plus since the advent of the cloud, companies are figuring out better ways to utilize offsite data storage and management. That often means avoiding a complete rebuild of an IT system on the cloud but using it in conjunction with on-premises data storage and functions. A hybrid cloud strategy should not be looked at as a threat to IT employee job security either, but as a means to free up valuable team members to focus on real-time changes and updates needed for data operations. Utilizing the cloud as part of a hybrid strategy will not eliminate IT jobs.
The future for many businesses is a hybrid cloud strategy, a mix of on-premises server and storage systems, coupled with offsite cloud partners that may be hundreds of miles or more away. When used properly as part of a hybrid strategy, it can free up talented IT employees to tackle even bigger issues.
About the Writer
Samrat Aich is an IT professional and solutions architect with 17+ years of experience and a proven track record of driving projects from start to finish in enterprise IT Systems, storage, virtualization and Cloud computing and Cloud security. For further information, please email samrataich [at] gmail [dot] com
The post Two Approaches to a Hybrid Cloud Strategy first appeared on IEEE Computer Society.
0 Commentaires