Bring 'em home: Repatriate cloud workloads, the HCI way
(2020/04/07)
- Reference: 1586278807
- News link: https://www.theregister.co.uk/2020/04/07/bring_cloud_workloads_home_hci/
- Source link:
Sponsored Cloud computing has become an everyday part of the mainstream IT landscape. It has even become a common view among some industry experts that all workloads will inevitably migrate to public cloud platforms, and that any organisations not working towards this goal risks being left stranded.
In reality, the benefits of moving to the cloud are not always so clear cut. It can be notoriously difficult to accurately estimate the long-term running costs of a cloud project, precisely because of the pay-as-you-go billing model for using resources. For this reason and others, many adopters of cloud services have been opting to shift some workloads back on-premises again.
This is a trend dubbed “cloud repatriation” by analyst IDC. A recent [1]Vanson Bourne survey of 2,650 IT decision makers found that 73 per cent are migrating applications back from the cloud to on-premises infrastructure. One notable example of this trend is Dropbox, which disclosed the same year it had saved $75m by migrating everything back from the cloud to its own data centres.
In addition to growing costs, respondents in the IDC survey cited security concerns, regulatory compliance, application performance and maintaining control over applications and data as the biggest drivers for cloud repatriation.
This does not mean that companies are turning their backs on the public cloud – far from it. Gartner predicts the public cloud services market will grow by 17 per cent during 2020. What it does mean is many are waking up to the idea one size does not fit all situations – that a hybrid model with a mix of on-premises IT complemented by services delivered from the cloud is the optimal approach.
Those repatriating workloads are not simply going back to the old ways of buying and running tin. Having experienced the benefits of cloud, they want these benefits – just on prem. What benefits? The ability to provision resources on demand via some kind of self-service portal, to scale resources up or down as required, to run a charge-back model so users can track the resources they are consuming and, of course, the cost.
Delivering an on-premises cloud
For IT departments, these requirements are something of a challenge. It can be extremely difficult to pull together an existing menagerie of servers, storage and networking gear into the kind of flexible, automated resource pool you would have experienced in the cloud.
This is where [2]hyperconverged infrastructure (HCI) comes in. Vendors first started offering ready integrated bundles of hardware known as converged infrastructure to address problems organisations were experiencing when building platforms capable of operating as [3]private cloud .
The converged systems market is growing, but it is HCI that’s leading the charge, according to [4]IDC . That’s thanks to HCI’s reduced complexity, ease of deployment and – significantly – what IDC calls “an excellent fit” within [5]hybrid cloud environments.
HCI goes further than converged by integrating everything – compute, storage, networking – into a single appliance-like box, which functions like an infrastructure building block: you scale out by adding more nodes. It offers the advantages of the cloud but without the risks associated with trusting your data to a third party.
Each node looks like an industry standard server, because that is essentially what they are. The key difference is that HCI relies on direct-attached internal storage rather than connecting to a Storage Area Network. The internal storage of each system makes up a shared pool of software-defined storage capacity across a cluster of nodes.
While early HCI systems were supplied as an appliance, or a pre-integrated unit, some HCI platforms are now also available as a software-only system that can be deployed on any industry standard hardware that meets the necessary specifications. That means organisations can continue to source systems from their preferred vendor.
The crux of HCI is that it offers greater simplicity and efficiency. Because resources such as compute and storage are virtualised and software-defined, they are largely automated; this makes provisioning much simpler than having to connect and configure a variety of equipment such as SAN storage arrays from different vendors.
The key piece of an HCI platform is that it should provide a unified management layer for the entire stack, enabling compute, storage and virtual machines to be controlled from a single console. It also has the big advantage of offering a single point of contact for support, should any problems surface. Maintenance releases and upgrades also apply to the entire stack, rather than IT staff having to deal with piecemeal updates for individual parts of the infrastructure, which may break compatibility with something else.
This can be seen in figures in an [6]ESG Research Insights Brief that indicate organisations using HCI saw a 25 per cent reduction in infrastructure operating costs, spent 31 per cent less time on routine system management, and saw a 52 per cent reduction in time spent on infrastructure deployment tasks.
Another feature of HCI is that services often delivered through separate, third-party applications can be built into the platform. These may include redundancy and high availability, backup and restore, and monitoring. Even if a platform does not have all of these capabilities to start with, HCI vendors typically add these with successive value-add updates of the software.
Going hybrid
For all that, there are use cases where public cloud is a good option, which has implications for IT teams moving workloads between public clouds and private infrastructures. They face another level of complexity when it comes to managing the resources from one or more public clouds in addition to those deployed on-premises, in what has become a hybrid IT ecosystem. As Gartner senior director analyst Ross Winser [7]points out : “Hybrid IT environments are becoming the new normal, with many enterprises managing workloads in data centres, in the public cloud and at the edge.”
Just one of the issues with this new hybrid IT ecosystem is the lack of consistency between on premises and public cloud platforms , especially in the management tools and skills needed to use them. This can end up creating yet another group of silos. IT managers would obviously like to treat all their resources as a single platform, with the ability to deploy applications anywhere, and move them around if necessary.
One answer that HCI vendors and cloud providers have both been working on, is to provide a consistent environment by replicating the HCI platform onto one or more public clouds, thereby enabling it as a cloud service. This is achieved by packaging the HCI software so that it can operate using the cloud as its infrastructure layer, instead of running on hardware in the customer’s data centre.
Running an HCI platform in the public cloud may seem counterintuitive, but it has advantages. For one thing, it becomes simpler to shift a workload from on-premises HCI to public and vice versa, making the process of cloud repatriation less costly and complex.
It also means workloads can scale beyond the capacity of the organisation’s on-premises infrastructure using resources of the public cloud while being managed by an HCI platform. This capability, dubbed cloud bursting, has long been talked about in reference to public cloud services, but has rarely been practical in reality.
Why? Thanks to various issues such as interoperability between the platforms at either end, which can be exacerbated by the habit of cloud providers regularly updating their services and APIs creating a race to keep up. Having access to a hosted version of the HCI platform can insulate you against this by providing a consistent environment that acts as an abstraction layer.
Of course there’s a catch and it comes from the fact that the HCI platform needs management tools that are able to coordinate both the private and public cloud sides of infrastructure through the same console. Currently, one [8]HCI platform allows for compute clusters running on a public cloud to be put into hibernation in cloud storage should they not be required and re-activated at a later date when they are needed, sidestepping the need to recreate them from scratch. This kind of feature lets you burst to cloud without incurring the ongoing expense of keeping machine instances operating and incurring compute costs, even when they are not processing data.
Cloud might be growing but on-premises infrastructure is enjoying a resurgence and HCI is in rude health. If it’s the technology benefits of public cloud you want but without the challenges or the costs, then HCI may be the way forward. But this is not an 'either' or 'choice' between public vs. private; rather, HCI can also serve as the bridge between the two, thereby enabling your hybrid-cloud environment in this re-aligning world.
Sponsored by Nutanix.
Sponsored: [9]How To Accelerate Brilliant Digital Experiences With Low-Code
[1] https://go.theregister.co.uk/k/Nutanix_enterprise-cloud-index
[2] https://go.theregister.co.uk/k/Nutanix_hci_for_dummies
[3] https://go.theregister.co.uk/k/Nutanix_top-20-private-cloud-questions
[4] https://go.theregister.co.uk/k/IDC_Worldwide_Converged_Systems_Market
[5] https://go.theregister.co.uk/k/Nutanix_designing-and-building-a-hybrid-cloud
[6] https://go.theregister.co.uk/k/Nutanix_ESG_The_Role_of_Converged
[7] https://go.theregister.co.uk/k/gartner-top-10-trends-impacting-infrastructure-operations
[8] https://go.theregister.co.uk/k/nutanix_hci_product
[9] https://go.theregister.co.uk/tl/1936/-8552/how-to-accelerate-brilliant-digital-experiences-with-low-code?td=wptl1936
In reality, the benefits of moving to the cloud are not always so clear cut. It can be notoriously difficult to accurately estimate the long-term running costs of a cloud project, precisely because of the pay-as-you-go billing model for using resources. For this reason and others, many adopters of cloud services have been opting to shift some workloads back on-premises again.
This is a trend dubbed “cloud repatriation” by analyst IDC. A recent [1]Vanson Bourne survey of 2,650 IT decision makers found that 73 per cent are migrating applications back from the cloud to on-premises infrastructure. One notable example of this trend is Dropbox, which disclosed the same year it had saved $75m by migrating everything back from the cloud to its own data centres.
In addition to growing costs, respondents in the IDC survey cited security concerns, regulatory compliance, application performance and maintaining control over applications and data as the biggest drivers for cloud repatriation.
This does not mean that companies are turning their backs on the public cloud – far from it. Gartner predicts the public cloud services market will grow by 17 per cent during 2020. What it does mean is many are waking up to the idea one size does not fit all situations – that a hybrid model with a mix of on-premises IT complemented by services delivered from the cloud is the optimal approach.
Those repatriating workloads are not simply going back to the old ways of buying and running tin. Having experienced the benefits of cloud, they want these benefits – just on prem. What benefits? The ability to provision resources on demand via some kind of self-service portal, to scale resources up or down as required, to run a charge-back model so users can track the resources they are consuming and, of course, the cost.
Delivering an on-premises cloud
For IT departments, these requirements are something of a challenge. It can be extremely difficult to pull together an existing menagerie of servers, storage and networking gear into the kind of flexible, automated resource pool you would have experienced in the cloud.
This is where [2]hyperconverged infrastructure (HCI) comes in. Vendors first started offering ready integrated bundles of hardware known as converged infrastructure to address problems organisations were experiencing when building platforms capable of operating as [3]private cloud .
The converged systems market is growing, but it is HCI that’s leading the charge, according to [4]IDC . That’s thanks to HCI’s reduced complexity, ease of deployment and – significantly – what IDC calls “an excellent fit” within [5]hybrid cloud environments.
HCI goes further than converged by integrating everything – compute, storage, networking – into a single appliance-like box, which functions like an infrastructure building block: you scale out by adding more nodes. It offers the advantages of the cloud but without the risks associated with trusting your data to a third party.
Each node looks like an industry standard server, because that is essentially what they are. The key difference is that HCI relies on direct-attached internal storage rather than connecting to a Storage Area Network. The internal storage of each system makes up a shared pool of software-defined storage capacity across a cluster of nodes.
While early HCI systems were supplied as an appliance, or a pre-integrated unit, some HCI platforms are now also available as a software-only system that can be deployed on any industry standard hardware that meets the necessary specifications. That means organisations can continue to source systems from their preferred vendor.
The crux of HCI is that it offers greater simplicity and efficiency. Because resources such as compute and storage are virtualised and software-defined, they are largely automated; this makes provisioning much simpler than having to connect and configure a variety of equipment such as SAN storage arrays from different vendors.
The key piece of an HCI platform is that it should provide a unified management layer for the entire stack, enabling compute, storage and virtual machines to be controlled from a single console. It also has the big advantage of offering a single point of contact for support, should any problems surface. Maintenance releases and upgrades also apply to the entire stack, rather than IT staff having to deal with piecemeal updates for individual parts of the infrastructure, which may break compatibility with something else.
This can be seen in figures in an [6]ESG Research Insights Brief that indicate organisations using HCI saw a 25 per cent reduction in infrastructure operating costs, spent 31 per cent less time on routine system management, and saw a 52 per cent reduction in time spent on infrastructure deployment tasks.
Another feature of HCI is that services often delivered through separate, third-party applications can be built into the platform. These may include redundancy and high availability, backup and restore, and monitoring. Even if a platform does not have all of these capabilities to start with, HCI vendors typically add these with successive value-add updates of the software.
Going hybrid
For all that, there are use cases where public cloud is a good option, which has implications for IT teams moving workloads between public clouds and private infrastructures. They face another level of complexity when it comes to managing the resources from one or more public clouds in addition to those deployed on-premises, in what has become a hybrid IT ecosystem. As Gartner senior director analyst Ross Winser [7]points out : “Hybrid IT environments are becoming the new normal, with many enterprises managing workloads in data centres, in the public cloud and at the edge.”
Just one of the issues with this new hybrid IT ecosystem is the lack of consistency between on premises and public cloud platforms , especially in the management tools and skills needed to use them. This can end up creating yet another group of silos. IT managers would obviously like to treat all their resources as a single platform, with the ability to deploy applications anywhere, and move them around if necessary.
One answer that HCI vendors and cloud providers have both been working on, is to provide a consistent environment by replicating the HCI platform onto one or more public clouds, thereby enabling it as a cloud service. This is achieved by packaging the HCI software so that it can operate using the cloud as its infrastructure layer, instead of running on hardware in the customer’s data centre.
Running an HCI platform in the public cloud may seem counterintuitive, but it has advantages. For one thing, it becomes simpler to shift a workload from on-premises HCI to public and vice versa, making the process of cloud repatriation less costly and complex.
It also means workloads can scale beyond the capacity of the organisation’s on-premises infrastructure using resources of the public cloud while being managed by an HCI platform. This capability, dubbed cloud bursting, has long been talked about in reference to public cloud services, but has rarely been practical in reality.
Why? Thanks to various issues such as interoperability between the platforms at either end, which can be exacerbated by the habit of cloud providers regularly updating their services and APIs creating a race to keep up. Having access to a hosted version of the HCI platform can insulate you against this by providing a consistent environment that acts as an abstraction layer.
Of course there’s a catch and it comes from the fact that the HCI platform needs management tools that are able to coordinate both the private and public cloud sides of infrastructure through the same console. Currently, one [8]HCI platform allows for compute clusters running on a public cloud to be put into hibernation in cloud storage should they not be required and re-activated at a later date when they are needed, sidestepping the need to recreate them from scratch. This kind of feature lets you burst to cloud without incurring the ongoing expense of keeping machine instances operating and incurring compute costs, even when they are not processing data.
Cloud might be growing but on-premises infrastructure is enjoying a resurgence and HCI is in rude health. If it’s the technology benefits of public cloud you want but without the challenges or the costs, then HCI may be the way forward. But this is not an 'either' or 'choice' between public vs. private; rather, HCI can also serve as the bridge between the two, thereby enabling your hybrid-cloud environment in this re-aligning world.
Sponsored by Nutanix.
Sponsored: [9]How To Accelerate Brilliant Digital Experiences With Low-Code
[1] https://go.theregister.co.uk/k/Nutanix_enterprise-cloud-index
[2] https://go.theregister.co.uk/k/Nutanix_hci_for_dummies
[3] https://go.theregister.co.uk/k/Nutanix_top-20-private-cloud-questions
[4] https://go.theregister.co.uk/k/IDC_Worldwide_Converged_Systems_Market
[5] https://go.theregister.co.uk/k/Nutanix_designing-and-building-a-hybrid-cloud
[6] https://go.theregister.co.uk/k/Nutanix_ESG_The_Role_of_Converged
[7] https://go.theregister.co.uk/k/gartner-top-10-trends-impacting-infrastructure-operations
[8] https://go.theregister.co.uk/k/nutanix_hci_product
[9] https://go.theregister.co.uk/tl/1936/-8552/how-to-accelerate-brilliant-digital-experiences-with-low-code?td=wptl1936