The Cloud Experience Everywhere

How HPE Metering Technology Delivers a Better Way to Consume IT

HPE GreenLake is a true pay-as-you-go, consumption-based approach to hybrid IT. It’s different than anything else on the market, so it’s not surprising that customers tend to have lots of questions. Here are some that I hear quite often:

Do I have to purchase the infrastructure?

No. HPE GreenLake is an “as-a-service” model, so we own and manage the equipment for you at your site, and we provide a buffer of capacity right-sized to your business. Then we measure how much you use, and we charge based on that usage.

What if I run out of capacity?

With metering and active capacity management, we ensure that there’s always a buffer of capacity available. We help you keep an eye on your usage, and we can always turn on more when it’s needed.

How long do I have to wait to get more capacity?

You don’t have to wait. We're always going to be ready to provide more capacity as you need it, because it’s already there onsite. You don’t have to wait for somebody to deliver boxes to your data center.

The key to understanding how it all works is understanding the metering technologies that power HPE GreenLake, enabling us to deliver a different kind of IT: on-premises IT as a Service. (See Don Randall’s post Defining the Next Chapter for the IT Industry.) We’re all familiar with the principle that to manage something, you have to measure it. HPE GreenLake’s metering technologies make measuring the usage of on-prem infrastructure – and therefore the ability to plan and manage capacity – a practical reality.

image_5 Things articles_v2_metering.jpg


How HPE metering technology works

Metering is the foundation of accurate billing and capacity management, so we tailor our metering process to each service that we offer. We do this by installing lightweight scripts that reach out to the equipment that we’re metering to pull the data on a regular basis. The system encrypts the data and sends it back to HPE where we process the data and feed it into the HPE Consumption Analytics Portal (see my blog Consumption-Based IT Just Got Even Better).

The portal gives you one place to go for visibility into usage and costs for both HPE GreenLake and public cloud infrastructure. It helps you plan capacity based on actual and forecast usage, and it enables you to create reports and dashboards for your user groups. We work with you to specify the parameters you’re interested in for the portal, and we apply sophisticated normalization and data cleansing technologies to the data to make sure you get meaningful reports for whatever dimensions you’re looking at.

That’s basically how it works, though the details may change depending on the solution you choose. For storage, for example, we measure the gigabytes (or other units that you specify) used on a particular device or set of devices. We look at what’s used every day of a given month and take an average, and we bill based on that.

For server usage, we can meter at the unit level – individual blades, for example – or at a more granular level, based on VMs. With an HPE Superdome Flex server, for example, you're likely more interested in tracking the VMs that run on top of it rather than the high-performance compute itself.

For VMs, we might take hourly snapshots of memory usage, and we track that usage by VM, which is a level that’s meaningful for doing showback of cost to a department or project. And if you have HPE GreenLake infrastructure in multiple locations, you can see what’s being consumed in each location.

Metering works a bit differently for HPE GreenLake workload/application solutions, such as Hybrid Cloud, SAP HANA, and Database with EDB Postgres (see the full portfolio here). These platform offerings include third-party software and more services, and each solution is unique. We generally use multiple units of measure. For example, with HPE GreenLake for backup, we meter the backend terabytes (i.e., how much you back up), and we also meter the licenses for the backup software itself, so you pay only for the hardware and software that you’re using.

HPE GreenLake gives you secure, transparent control of your data. Your usage and costs are always available for you to view in the portal, and you have access to historical data to spot trends and help plan for future capacity. As well, you can set budgets, create reports, and build rules-based insights for any of the services we meter, to ensure greater control. These are just some of the important business benefits that sophisticated metering enables. It’s a key differentiator to HPE’s as-a-service offering.

Usage metering makes HPE GreenLake unique

Now that I’ve described what HPE metering technology is, how it works and what it does, I’ll briefly cover what it’s not, and why HPE GreenLake is very different from, and much better than, what other vendors are offering.

First off, it’s important not to confuse this as-a-Service model with leasing plans. The idea of renting IT infrastructure has been around forever, of course, and some vendors are eager to rebrand it as consumption-based IT. But these are two very different beasts. Leasing does enable you to spread your costs over a period, but because it lacks metering capabilities, it doesn’t let you track actual usage of the assets. So you could still end up paying for capacity that you don’t need and don’t use. And if you want to ramp up capacity, you’re stuck with the usual wait times, just as you would be with a Capex model.

Similar capacity management issues arise with other metering models announced by other vendors earlier this year, such as one based on power consumption – kilowatts per hour. It only takes a moment’s thought to see that there’s a huge difference between measuring the power that an IT resource is consuming versus metering actual usage of that resource. Looking at power consumption doesn’t tell you whether a server is using most of its CPU or memory, or whether a storage array is filling up.

That’s a crucial difference when it comes to planning and managing capacity. For one thing, it makes it really hard to predict your costs. You may know how many CPUs or gigabytes of storage you’re going to need for a project, but how do you translate that into kilowatts per hour? Plus, measuring power consumption can’t provide anything like the granularity you need to tie your usage and costs to a project or a part of your business. It’s more a billing model than a true capacity management capability.

Transparent, accurate, agile provisioning

Metering is what enables HPE GreenLake to manage your capacity so that there's just enough buffer for your usage patterns. Most businesses overprovision by a lot, often around 50 or 60 percent. With HPE GreenLake, you don't have to fill up your data center with boxes that you don't think you'll ever use, just to be on the safe side. We show you your usage patterns and make sure you stay ahead of resource demands by installing just the capacity you need, when you need it. Effectively, you’re getting 100% economic utilization of your valuable IT resources.

You're saving the money you would have spent on overprovisioning. You're also saving on maintenance in the data center and reducing your data center footprint. And you can get new projects up and running without a lot of upfront expense or waiting around for infrastructure deliveries and installations.

Read more about how you can gain cloud flexibility and on-premises control with HPE GreenLake.

0 Kudos
About the Author