Successfully Navigating Hybrid Cloud Challenges

According to a recent report from 451 Research, the ability to migrate workloads easily between on-prem and public cloud is the number one reason that organisations opt for a hybrid IT infrastructure. Unfortunately, a significant divide exists today between traditional enterprise IT environments and the public cloud with different management models, consumption models, application architectures, and storage and data services. These differences can limit your ability to easily move enterprise and cloud-native applications where you need them. By Peter Gadd, VP EMEA Core, Pure Storage.

  • 4 years ago Posted in

To create an effective, modern hybrid IT infrastructure, you’ll need to identify the differences that create the biggest challenges to your operations and take steps to close those gaps.

 

Performance

Most enterprises have dozens, even hundreds, of traditional enterprise applications, and there are many reasons to move some of these applications into the public cloud. However, many IT teams are unhappy with the results when they just “lift and shift” applications into a public cloud. Performance is often poor. Cloud-native applications achieve performance by scaling out, but enterprise applications more often scale-up—creating I/O performance requirements that clouds may not deliver. The public cloud may also lack the resiliency that enterprise applications expect.

 

Fortunately, a number of block and file data services are now available to support enterprise workloads in the cloud. These services layer enterprise capabilities on top of existing cloud data stores, offering pay-as-you-go consumption while delivering the high availability and efficiency your applications expect so that applications don’t have to be redesigned. They may also offer snapshots, cloning, and replication to support data protection, disaster recovery, and other management functions.

 

On the other hand, there are many applications that would benefit from the functional and economic benefits of the cloud but must physically remain on-premises (example: law enforcement video feeds).

 

Here, an On-Premises-as-a-Service (OPaaS) solution would align perfectly. When selecting a vendor, find one that supports an OPaaS model. The ability to utilise storage-as-a-service within your secure data center provides nearly all of the public cloud benefits, while retaining the security necessary.

 

Management

Management interfaces and tools in the public cloud are different than the ones you use in your data center, and there can be a significant learning curve. Administrators who have to switch back and forth between environments are likely to be less efficient. This adds to the burden on your IT team and increases the risk of errors. At a minimum, your existing administrators are going to need to be trained in both paradigms, and you may end up needing one team for on-premises management and another for the cloud, adding operating expense.

 

By choosing enterprise cloud data services compatible with the storage you use on-premises, you can provide consistent management interfaces across both environments. This unifies management tasks and makes it possible for the same administrators to operate effectively in both environments, reducing the risk of operator error and eliminating the need for a separate team for cloud admin.

 

APIs

Traditional applications that have been written to make API calls to your on-premises storage systems won’t run in the cloud where, in all likelihood, those APIs aren’t supported. Porting these applications to run in the cloud can take considerable time, expense, and developer retraining.

 

By choosing enterprise cloud data services compatible with the storage you use on-premises, you can ensure that APIs are consistent across both environments. As a result, applications written to those APIs will be able to run in either environment without porting, freeing your IT and development teams to focus on more important problems.

 

Storage

Object storage is widely used in the public cloud. Nearly 70 percent of the scale-out/scale-up storage market is object-based, according to IDC estimates. As a result, most cloud-native applications are written to use object storage. If you want or need to run those applications in your data center, you have two choices:

 

·         Undertake a substantial porting effort to enable the application to use file or block storage

·         Provide object storage on-premises

 

For many enterprises, the second option—on-premises object storage―is ultimately a better choice. Cloud-native applications are likely to be a bigger part of your company’s digital strategy. With cloud-compatible object storage in your data center, you’ll be able to develop once and deploy in any cloud. In addition to its role providing cloud-native application compatibility and portability, object storage is increasingly being used in the enterprise to support big data analytics, content management, and various backup and archival use cases.

 

Data protection and disaster recovery

For many enterprises, the first hybrid cloud use case is often to add cloud data protection and disaster recovery (DR) for enterprise applications and data. However, while it sounds like a no-brainer, putting all the pieces together to make it possible—let alone efficient and cost effective—can be a challenge.

 

Choosing primary and secondary storage that is “cloud capable” is the simplest path to cloud data protection and DR. Ease of cloud integration should be one of your selection criteria any time you refresh data center storage. Look for storage solutions that don’t require you to integrate a lot of third-party software or extra hardware. Of course, the reality is that most data centers continue to have at least some degree of heterogeneity.

 

An alternative solution is to add a cloud-capable secondary storage system as a backup target, backup all the old data to that, and then replicate the results to the cloud for retention and DR.

 

Infrastructure

As with most decisions, the answer is not binary. The correct hybrid infrastructure is typically a mix of on-premises, collocated/hosted, “connected to” the public cloud and “within” the public cloud. To this end, most users would prefer a unified infrastructure that provides a common user interface, a common cloud experience and a common subscription/license—allowing them to align application workloads with the most appropriate hybrid cloud infrastructure.

 

Look for vendors that can not only support a variety of hybrid options but can provide a unified/common experience and mobility between the configurations. This way you can’t make a mistake because you can tune your hybrid cloud by moving your workloads as needed.

 

Addressing these criteria will enable you to erase the cloud divide and build a modern IT environment that delivers the best-case scenario for data protection, performance optimisation, cost optimisation and business agility.

 

- ENDS - 

 

 

 

By Tytus Kurek, Product Manager, Canonical.
It is an unavoidable fact, high performance computing (HPC) is an energy intensive environment....
We live in a world that is increasingly driven by technology. As we continue to increase the number...
AI has begun to change many facets of our lives, creating tremendous societal advancements. From...
Greater performance, less capital expenditure, and lower operating costs with AMD EPYC processors.
The manufacturing industry is producing more data than ever before as sensors become readily...
The internet has become a fundamental part of so many businesses around the world. A large...
In 2020, we are used to the idea that physical goods imports include a degree of embodied carbon,...