- About Us
- News and Events
- Contact Us
Watch this webinar to understand how virtualization effects Oracle licensing both on-premise and in the cloud, and how to leverage virtualization for your business.
We’re going to begin right here. Really, this is about virtualization and Oracle in the Cloud Age, and how Oracle has gotten to where it is with its positions, and what those positions are relative to not only VMware, but other hypervisors, including Oracle Virtual Machine. We’ll also touch on the Trusted Partitioning Policy, which is the guiding document that Oracle uses. We’ll talk about some of the other tools that are available to folks to help segregate their environments in a virtualized manner.
Just one thing to get out of the way, basic housekeeping. Really, this presentation is for educational purposes only. Your licenses that you own from Oracle, the rights to them, are governed by the contractual agreement that you have spelled out in your Oracle License and Service Agreement, the Oracle processor policy, and the Oracle Software Technical Services Support Policies. The presentation you’re about to be presented is for general background information on common use cases and configurations that may not be perfectly applicable to your situation. In essence, what that means is everything you’re getting here is an overview of an interpretation of the policy.
The agenda is really cut and dry. We’re going to understand virtualization effects on Oracle Licensing, both on-prem and in the cloud. We’re going to talk about various Oracle licensing documents and terms, the architectural factor, virtualization partitioning, VMware through the ages, the Oracle Partitioning Policy which is the Oracle’s governing document. We will touch on Nutanix and a little Q&A at the end.
So, really, if you think about it, over time, companies are forced to choose, really always, between flexibility or lower cost. When you look at things, virtualization is really a key component in there, right? You want to be able to rebalance your spend across different services. That’s a difficult thing. The visibility or control over software spend really is limited. Now, in the Cloud Age, it’s become even more difficult to understand and to monitor. Companies really have difficulty leveraging their on-prem software investments in the cloud and in this hybrid cloud environment, if some of you are in it. Or if you’re still in a traditional on-prem environment, it’s still very difficult to manage that. The last point here, which is delayed time to value because of complex pricing, that’s always been an issue with Oracle and continues to be.
So, the first thing you need to think about here is knowledge of Oracle’s Virtualization Policies and Best Practices really can make your life very simple; can avoid the trappings that lead to major unbudgeted, unplanned expenditures; and can make it more affordable for customers to move to the cloud while creating extreme flexibility in what and how they use cloud services.
So, one of the key things to bear in mind is Oracle’s Processing Core Factor, and we’ll talk about that a number of times throughout this presentation. For those of you who don’t know, it’s a simple standard that Oracle has applied based on chip set and manufacturing. If you look on the screen, you’ll see there’s a link that you can Google or you can put it in your browser, and it shows you exactly what the core factor is, whether it’s 0.5-to-1, which is half an Oracle processing license per licensable core in an Intel type chip, to a one-to-one processing factor if you’re in an IBM power type chip. So these core factors play not only in your on-prem environment, but also carry through to the cloud, and we will touch on that later.
When we’re talking about virtualization … and you’ve heard me mention Oracle’s Partitioning Policy, and we will get to that, and there’s a link that you’ll see later on in here. That’s really a page-and-a-half document that Oracle has published which is outside of your software license agreement. It’s out there, and it’s the positioning document that Oracle leverages, both from a sales perspective and from a business practice or license management systems perspective, LMS, as to how your licenses are governed. In that document, there is a section that talks about soft partitioning, another section that talks about hard partitioning, and about trusted partitioning. These are three different turns.
So soft partitioning would be virtually carving out a processor. Hard partitioning is segregating out processors from the whole to isolate them in their own little center. Trusted partitioning is the ability to expand on demand with Oracle’s engineered systems, whether that be Database Supplies, whether that be Exadata, Exalytics, Exalogic, any of those types of machines. That leads itself to the on-demand licensing, which, again, is on the engineered systems that allows you to expand licensed consumption as needed so that you can go from, let’s say, 10 cores to 15 cores with just the [effectuazation] of enabling those processors and possessing those licenses. That is something that is only available on the Oracle engineered systems. We will be talking about the public cloud. You folks probably know names like AWS, Azure, Oracle—OCI is Oracle’s cloud—amongst others.
So, as we think about it, Oracle’s licensing has evolved throughout the ages. When Oracle began and it was out there, multiprocessors started taking what was a client server type environment and expanding on it. So Oracle went from a license set where you had a concurrent device type license, which were the number of users at any high water mark on a server or servers, to what they said was a core factor. In between, they did megahertz, but for this discussion, we’re going to focus on going from a core processor core factor, so you count the number of processors in a machine, and that would be the number of licenses you have.
However, the computer sprawl that occurred during that time, as companies expanded laterally and vertically, led to the rise of virtualization, because the sprawl became unmanageable, untenable, and very expensive. So organizations began to virtualize their machines. So taking a physical server and basically virtualizing it and cutting it into a virtual amount of forests. If you had a 12-processor machine, you could then cut it down or virtualize it into tens of processors on each individual core, thereby expanding your capability without creating digital sprawl, but it led to creating virtual sprawl. That, in turn, though, that rise of that technology led Oracle to develop and implement their partitioning policy, which basically is going to state that any processor that you virtualize, you have to license at the physical level.
However, as virtualization gets more sophisticated, that policy, being broadly and generally written, allows for the expansion of this policy to go beyond even the processors that are running Oracle, and we’ll talk about that. Then, we’re going to talk about hyper-converged infrastructure. The hyper-converged infrastructure is really the evolution of where we are now and leads into the cloud, which the cloud dynamically changes everything. We do see a matter of resources you need yet being policy virtualization, because some clouds are bare metal and some are virtualized, still is in effect. However, there are various degrees of effect that the partitioning policy has in the cloud, depending if it’s Oracle’s cloud or an AWS cloud. There’s a limit on it.
Just as a point of reference, if you’re looking at AWS, it’s running on x86 here. However, since the partitioning policies state that any core that could be accessed through a V-motion needs to be licensed, but because AWS is so expensive and Oracle has a custom agreement, in lieu of applying this policy, which would make the cloud untenable for any company to use, what Oracle instead did was say that any core inside of AWS that you’re using must be licensed at a one-to-one ratio, not at the 0.5-to-1 ratio that we’d normally apply to x86 or Intel-based chips. So that’s the concession that Oracle made to allow you to use the cloud in a virtualized manner without really having to license every single processor that’s in this cloud. So that’s just an example of how this partitioning policy kind of has the ability to morph itself to situations. It really needs a tight view to determine where and what you’re doing.
Although it seems very simple, every physical core that you virtualize needs to be licensed for Oracle license if you have it on there. However, any core that can be accessed through the vCenter also needs to be licensed, regardless of whether or not you’re running Oracle on it, which now that poses major challenges in the Cloud Age.
Really, as we look at it, here’s the definition. I’ve mentioned it a few times. This is the heart and meat of everything, right? It’s when CPU on a server is separated into individual sections where each section acts as a separate system. It can be called segmenting, but normally, it is partitioning is the term we use. There’s several hardware and software virtualization technologies available that allow partitioning abilities with varying degrees of resource allocation. Why do you do it? It allows you to perform resource consolidation, increase uptime, and decrease running costs.
Now, here’s the rub, right? Soft partitioning, again, segments the operating system using OS resource managers. The operating system limits the number of CPUs, right, where Oracle database is run. What does this really mean? If your solution uses software to create separate silos of compute on a single physical device, you’re using soft partitioning. That can be VMware, Hypervisor-V, AIX Workload Manager, Oracle Virtual Machine. Again, this is a repeating theme. Count all the physical cores in the machine, multiply by the core factor. However, you need to bear in mind what is part of the vCenter. So it’s not only what is on that machine, but what other machines are part of this sphere?
So if you had a machine that was running Oracle on it on 12 cores and you had four other cores on there that were running not Oracle, they’re on the same server. So in Oracle’s vernacular, unless it’s on a machine that allows for hard partitioning, you have to license all of those cores, regardless of the Oracle. Concurrently, if there were second and third machines also not running Oracle but in that same vCenter, all the cores on those machines would have to be licensed. So to recap this, really, the first thing you’re going to do is count all the physical cores that have Oracle on it. Then, you need to look at what and if you have other cores in there that could be VMotion, too—A, on the same physical server or, B, on other servers that are in the same vCenter.
In addition, and we’ll touch on this later on, with VMware, as it evolved over time, it had the ability to VMotion throughout. So shared storage will play a factor in it, and we’ll get to that shortly here. As we talk about Oracle’s partitioning categories, we talk about hard partitioning, and I mentioned that earlier to you. Hard partitioning is really physically segmenting a server. Most of you, if not all of you, understand that, where you can separate, at the boot area, memory input, and the like. From a licensing perspective, if your machine is the correct type, it is possible to physically partition cores, and the licensing required thereby is physically partitioned.
So in that example, again, as we talked about before, if you have a machine with 24 cores on there, but you can isolate and physically partition, through hard partitioning, six cores to be Oracle only, then those six cores would be the only ones requiring licensing. You would take those six cores and multiply them by the per core factor. However, if you can’t, that’s a different story. This is just in a bare metal type environment. When we talk about virtualization, even with the hard partitioning, if you can VMotion from the hard partitioned Oracle to the non-Oracle core, that becomes another problem. Okay? Remember, all approved hard partitioning technologies must have a cap or maximum number of core processors for the given partition. Key factor in that was Oracle.
Now, this is very interesting here. Oracle’s view on virtualization, if you think about this for a minute and you say to yourself, “Well, Oracle’s saying if I have a server and I’ve got 24 cores on it, I’m setting 12 up to Oracle. I virtualize those 12. But I have 12 non-Oracle cores on there, and I have two other servers that run non-Oracle all on the same vSAN.” Oracle is saying I’ve got to license every single physical core on the three machines. That’s akin to saying, there’s a parking lot. Every spot in the parking lot costs $100, but I’ve got 100,000 spots in here or 10,000 spots in the lot. Because I could park in any given spot in the lot, I have to pay for every single spot in the lot. That’s really the visual that we’re trying to impart here with this picture.
Oracle is saying because you could move to any core in the vCenter, you have to license all of them. That, in tune, goes to two factors. If you’re on VMware below 6.0 and you’ve got shared storage, the view is you’ve got to license everything that is in the same … has the ability to VMotion that’s on that same shared storage. However, if you have non-shared storage, and we’re going to get to this, you can isolate and limit the amount of spaces that you need or cores that you have to license. However, if you’re a VMware 6.0 or above, that becomes a non-factor with the shared storage, because that allows you to VMotion anywhere, regardless of the shared storage, which becomes a major cost factor as you think about virtualizing.
I can tell you, from first-hand experience, that we have seen clients who have small environments, small user populations, but didn’t really understand or architect correctly when they designed their virtualized infrastructure and were exposed to the tune of millions of dollars’ worth of licensing requirements, both for database and middleware Oracle licensing because of this. Remember, this policy that we’re talking about runs on what Oracle calls a technology stack, which is databasing tools and middleware. So if you, for instance, for a UBS customer and you’ve got database in WebLogic or Internet Application Server, this is not only on the database layer, but on the middleware, whether it’s WebLogic or Internet Application Server. That’s a major point of exposure here. That’s why we keep referring back, during this presentation, to exposure to risk.
Really, what we’re trying to bring out and make everyone aware of is what seems to be very simplistic and easy to manage can become something as catastrophic as what’s indicated on this picture. So, Oracle recognizes the practice pay for server usage based on the numb of CPUs that are turned on, capacity on demand or pay as you go. This feature, however, is available on some of the engineered systems, whether that’s the database appliance for the Exadata … which means that if you had an ODA with 24 cores on it and you began its service with six of the cores lit up and licensed, over time, you can grow this to any number up to the maximum capacity. Because it’s trusted, you also have the ability to run non-Oracle on different parts of that device and not worry about the fact that you are in the same unit and having to license all cores regardless of the Oracle.
So, again, on-demand licensing, the ability to pay as you grow, is limited to the Oracle engineered systems, and it has a twofold benefit on it, right? Pay as you go and trusted partitioning. OVM, which is the leading way that you would do it on one of those two systems that we talked about. OVM is Oracle’s Virtual Machine. It offers advanced features for hard partitioning, also known as CPU tinning. That’s regardless of what brand of hardware it’s on. As long as you can run it on there, it’s available. So it could run on Linux. It could run on Solaris Spark. You have the ability to do it. Hard partitioning means binding the vCPUs to a physical CPU thread or cores and preventing these vCPUs from being scheduled on physical CPU threads or cores other than the ones specified. It is application driven.
Unlike VMware, Oracle’s virtualization solutions know what is running inside the virtual machine and can provision and manage those application’s middleware and database and coordinate it. It is low cost. Unlike VMware, Oracle VM is free to download, use, and distribute, with extremely affordable support fees. So for those of you that are saying, “Why do it?” Well, A, it reduces license exposure. B, it is extremely economical, thereby reducing your cost of VMware. The skill set can be learned very rapidly, or you can use a partner to help you to implement it. Regardless, it is a very low-cost, fully functional method to virtualize. It’s very rapid to implement. The templates make it very quick to deploy enterprise software, including CRM, ERP clustering, and management. Oracle VM Reference Configurations accelerate deployment of complete virtualized infrastructure.
It’s got integrated management, so it’s got full stack management, which lets you effectively manage the Hypervisor physical server in an application through one single pane of glass. It’s a very efficient architecture. Having a single engineering team for Linux and OVM leads to better optimization between Hypervisor and Linux. Remember, these are both Oracle products. Although Linux is socialist software and it’s free, Oracle has its own flavor of its support, and it’s the same kernel as Red Hat. Again, less expensive than the Red Hat support. It is, again, free software. You only have to pay for the support. I know I’m melding two things together, but if you think about the OS and the virtualization layer and you say, “I’m running Oracle technology.” Why wouldn’t you run Oracle virtualization and Oracle Linux, thereby reducing costs, have a single pane of glass, and single neck to choke, so to speak? Makes things very, very cost effective, very easy to manage.
It is an extremely efficient method to reduce exposure to risk and reduce costs, all with the intent of managing your on-prem and giving you a good handle on sprawl. So, again, Oracle makes it easier and more affordable for customers to, A, get license mobility. OVM makes everything easier and more flexible and less expensive.
So we’re going to skip over the universal credits for cloud, because I’m not sure we’ve got enough time to talk about that, but we can get to that later. Essentially, you really have the ability to deal with virtualization through some different tools, right? The key points to remember is this. If you have short storage and you’re running VMware below 6.0, all cores in that vCenter that can be accessed through that shared storage have to be licensed for Oracle, regardless of whether or not they’re running Oracle. If, however, you can give dedicated storage to a dedicated Oracle vCenter, you can isolate the Oracle licensing required to those cores that are in that vCenter. If, however, you are on VMware 6.0 or above, really, all cores in your data center that could be accessed through vMotion, regardless of the underlying shared storage, will need to be licensed.
There is a major point of exposure when you think about what you have here. You need to look good and hard at your environment architecturally for what you’re doing. Just as a case in point, we had a client who was designing a warehouse management system, and they were trying to understand how best to architect that system. The problem they were running into was they were all on VMware 6.4. the number of cores in their environment are over 2,000 physical cores. What we said to them was, “You’re a non-Oracle environment for the predominance of these cores.” They were an SAP shop running Db2 for their environment, and they had non-Oracle throughout their VMware from Microsoft and other things. We said to them, “Your issue to put 24 cores for a warehouse management system is going to expose you to millions of dollars of Oracle licensing.”
We basically had to guide them through out to re-architect using Oracle Virtual Machine, VMware, and isolating out those cores onto the servers that were going to be dedicated for this warehouse management system. What they ended up with was a highly efficient, highly leveraged environment that did not require them to have all the Oracle licensing, but only license the 24 cores that were required to be licensed. In turn, again, that story highlights the fact that you need to pay attention to not only the cores that are running Oracle, but the cores that aren’t, and the underlying shared storage, which some people may or may not know about. So those are the key takeaways back and the Oracle Trusted Partitioning Document as the guiding emphasis of it.
So, what I’d like to do now is give you all a few minutes on Centroid, who we are, what we do. Then, once we’re done with that, we’re going to break into a quick question and answer session. For those of you who don’t know or aren’t familiar with Centroid, we are an Oracle Platinum partner, specifically a top 25 strategic partner. We span the gamut from servers and storage up through the OS, into the database layer, and all the way out into the front end ERP or edge applications like EBS, Hyperion, OID, and the like. We do everything from project-based solutions, customization, management services, hosting, cloud assessments, cloud migration, cloud monitoring, licensed assessments, licensed optimization, licensed management, and ongoing expense optimization.
We’ve been around for over 20 years. We’ve got offices throughout the US, from the New York metropolitan area to Dallas, to Detroit, to Northern/Southern California, to Hyderabad, India. We’ve got over 300 or so Oracle professionals that we have employed. Our focus is on future-proofing your environment to make your business more efficient, more economically run, and more cost-effective. Really, that’s who we are. Just a few things to note. We are an Oracle gold-level cloud partner and have implemented some of the largest infrastructure and platform implementations in the Oracle cloud. We also implement non-Oracle cloud, such as AWS and the like.
We are the Oracle engineered systems partner of the year for several years running, so we have extensive Exadata, ODA, Exalytics experience. As you can see, there are too many things for us to kind of cover in a nutshell, but we would gladly welcome conversations on any and all of these in the future, if you have any. Suffice it to say, I know everybody’s time is valuable, and I don’t want to kind of take this up with a commercial. You guys came here to learn stuff. So what I want to do is go over a few questions that have come in while the webinar was going on. What we’ve seen so far is we’ve seen a few questions come in as to what type of assessments are offered by Centroid to help reduce the risk of non-compliance through VMware?
That’s really an excellent question. I wanted to cover that in a little depth. One of the things that we encourage all of our clients to do, and Gartner and Forrester both also recommend doing this, is that on an annual or, at the very worst, every two years, to self-assess your entitlement rights against your actual deployment. What Centroid actual does to that effect is we actually assess the actual software license entitlement rights that you have in your original software agreement, such as your software purchases and the like. What we do is we look at them against how things were deployed in the past and how they’re deployed today and where you’re going in the future. Then, we give you a gap analysis.
If you’re virtualized, we look at the virtualization, and we then assess them against where you are and what, if any, gaps do exist. I know that there was a comment that came in from Sean. You wanted me to go over the other slide. I will do that with you on a one-on-one basis, because it’s going to be a little sloppy for me to exit where we are today. I’ll go over that with you offline, and I’ll have Bob set up a time to go over that in depth with you. I apologize for it, but I don’t want to come out of this to go to that right now, if that’s okay.
One of the other questions that came up here was, what really is the relationship of AWS to licensing rights? That’s a very simple question to answer. So you pay for your virtual cores on AWS by the core, right? So whatever you contract with AWS for, you need to license per core with AWS. So if you’re paying for 10, you need 10 Oracle licenses that are going on there. It’s a one-to-one ratio. However, if you are an Oracle cloud, that ratio is two virtual CPUs per one Oracle license. So there is a definite cost factor to go to that. So that is one area that is available to us to help reduce costs, is to look at the Oracle core and the Oracle virtual machine.
There’s another question here, and that question revolves around if you own a ULA, how does that play into virtualization? That’s an excellent question. The ULA, as I believe most of you probably know, is an Oracle Unlimited License Agreement. What that means is you get to deploy unlimited quantities, unless otherwise specified in your agreement. However, unless there is a preclusion written into that original software agreement, you can deploy an unlimited quantity in a virtualized environment. If, however, there are preclusions written in there, it would preclude you either in the cloud or on-prem in a virtualized environment. So it’s imperative that, again, you go back to your Oracle documents and look at them closely to ensure that you are out of or in compliance, whichever way it is, and that you’re following the guidelines in there. Again, the original software purchase agreement that you have with Oracle or through the partner that you bought them through are the governing documents superseded by the Oracle Partitioning Document.
Well, we’re coming up against 11:45 Eastern, and I want to be mindful of everybody’s time. If there are any other questions, please feel free to email them to Bob Huggins. That’s [email protected] Or to me, [email protected] I thank you again for your time this morning, and we look forward to meeting with you again on another webinar or in person at Oracle OpenWorld, where we will be having an event Sunday night. It is a cruise in the harbor, so we encourage you, again, to reach out to us, let us know that you’re attending. We will subsequently get you an invite onto the vessel. So, again, thank you, and take care.
Centroid is a cloud services and technology company that provides Oracle enterprise workload consulting and managed services across Oracle, Azure, Amazon, Google, and private cloud. From applications to technology to infrastructure, Centroid’s depth of Oracle expertise and breadth of cloud capabilities helps clients modernize, transform, and grow their business to the next level.