Hello everyone, welcome to the launch of Oracle Solaris 11.1. My name is Markus Flierl. I run the Solaris Core Technology Group here at Oracle, I'd like to give you an update on what we actually released with Oracle Solaris 11.1 and also highlight a little bit about the actual customers that we already have today, what customers are already doing today in their real-world environments with Oracle Solaris 11, how Oracle Solaris 11 is allowing them to transform their IT environment, how it's allowing them to innovate, how it's allowing them to become more productive in their environments. Before I jump into the details, one of the big things that's really changed for us in the Solaris organization has been the increased investment that Oracle has been putting into us. My team, for instance, has grown by about 30% in the last two years.
We've had a huge influx of new talent, people coming, both on the experienced side as well as college hires, that are helping us to actually drive the technology forward, both in terms of some of the features in Solaris, also working with some of the other teams around Oracle to figure out how we can integrate things and provide customers with the best possible overall solution. Before I go into the details about Solaris, I just want to quickly highlight the bigger picture and the bigger strategy of Oracle. For those of you who tuned into Oracle OpenWorld a couple of weeks ago, you've probably seen the massive investments that Oracle is making in the cloud space. What this means for a customer is that, if you want to run in Oracle's public cloud environment, we can provide that for you.
If you want to run those kind of applications that we have in the public cloud in your private environment, but have it fully managed by Oracle, we can provide that. If that means you want to have a hybrid of the two, we can provide that for you. It doesn't matter whether this is software as a service, platform as a service, or infrastructure as a service, we provide all that in a very comprehensive fashion. Not only do we provide that, but we also give you a very flexible path for adopting that for your environment at your own pace, depending on what your situation is. There's a whole combination of products that provide the foundation for these kind of capabilities, and Oracle Solaris 11 is a key one within that. We coined Oracle Solaris 11 as the first cloud operating system last year.
A lot of you are probably asking, why is that? What is it that is in there that is allowing us to do that? There's a massive amount of investment that we have made in Oracle Solaris 11 that we've continued with Oracle Solaris 11.1, in order to provide you with the capabilities to run your environment in the most secure and in the most effective way within your cloud. The two things that we're bringing together here is the ability to have all these traditional properties that Oracle Solaris 11 offers, security, availability, the observability with DTrace, all these other pieces, bringing that together with the latest and the greatest cloud infrastructure features, cloud management features.
Things like the automated install that we already had in Oracle Solaris 11, the fast and fail-safe packaging, the new IPS packaging system, all the investment that we've put into the virtualization, both from an OS perspective, also from a storage and a network perspective. Of course, providing you with the right tools around that such that you can actually manage this kind of environment at a very large scale, in the most safe, in the most secure, in the most available way by bringing these two areas together. Let me just show you some examples here of what customers are doing today. In the following, some customers obviously want to do infrastructure as a service. The example here is a major U.S. mobile provider. It's U.S. Cellular.
They're taking Oracle Solaris 11, they're running it on top of logical domains with the latest SPARC T4 processors, they're providing infrastructure as a service to their internal customers at a very attractive price performance point, highly available, highly secure. I'll be talking about that in some more detail later on. Other customers, they want to build a platform as a service. Here's an example of AAPT, which is a major Australian service provider, they're doing this today using the Oracle Database. They're using Oracle Solaris, they're using Oracle Solaris Zones. The nice thing about Oracle Solaris Zones is that it works across different hardware platforms. They can provide these kind of services on top of the Oracle Server X-series, the X86, the SPARC T-series, also the SPARC M-series products. Of course, other customers want to provide the full stack.
They want to provide the applications on top of the platform, on top of the infrastructure. We have people like Ativas, which is a Brazilian service provider that is doing that. They're providing SAP as a service in conjunction with the Oracle Database, in this case, running on a SPARC SuperCluster, which has all the optimizations for the Oracle Database, the latest enhancement in Oracle Solaris 11, it brings it together with the latest enhancements in the SPARC T4 hardware. All of this is running over 40 gigabit InfiniBand networking. Again, from what we've been talking about Oracle Solaris 11 being the first cloud operating system, for a lot of customers, that is already reality. They've already gone out they're already deploying the cloud infrastructures in this kind of fashion.
Yeah, at this point, these are just some of the highlights here in terms of some of the logos of customers that are doing interesting things today. There's already more than 2,000 customers already in production today. With the release of Oracle Solaris 11.1, I would expect significant uptake that we'll be seeing there. The other critical thing, of course, for customers, what customers need in order to be able to deploy it is having all the major ISV applications certified. At this point, we have most of the major ISV applications certified on Oracle Solaris 11. SAP, for instance, was certified within a few weeks of the actual launch. Most of the other major vendors like SAS, any other third-party vendors here are at this point on Oracle Solaris 11. Of course, there's a long list of Oracle applications.
Being part of the biggest ISV in the world now has certainly changed things around quite a bit, it's made it a lot easier for us to get a lot of these Oracle applications certified. In fact, some of the Oracle applications, like the Oracle Database and WebLogic, were already certified even before Solaris 11 shipped in the context of these engineered systems, in that context, they were actually running on Solaris 11 Express. Yeah. At a high level, what are we providing here with 11.1? It's very similar at a high level what we already did with 11, we're just building on top of that, which is continuing to provide the best cloud infrastructure. I will talk in some more detail about what that means.
Allowing customers to unleash the power of cloud, of virtualization that's underneath there, being able to become a lot more flexible, a lot more agile, quickly deploy services, quickly tear them down, quickly reprovision without having to worry about the actual underlying physical hardware. Then the other big area of investment, of course, is the integration up and down the stack. Traditionally, of course, we've been working closely with our partners on the hardware side. Being part of Oracle now, it allows us to work very closely with the various teams within Oracle, be it the Oracle Database, be it the Java team, be it other teams. We're working very closely with them to provide the best solution for the Oracle stack. Last but not least, as I always jokingly say to customers, it's still safe to run Solaris 11 in an on-cloud environment.
There's a lot of investment that's going into the traditional side of the house, where we're looking at performance, scalability, security, all these other kind of parameters that customers are used to. We are, of course, very heavily investing in those and making sure that we keep up with the latest innovations also in the hardware with bigger and better hardware that's coming down the pipe. In terms of the actual features, what does this really come down to? What does this actually provide for a customer? What do we mean with this lifecycle management in the cloud with the cloud infrastructure features that we have in Solaris? The slide here shows you the typical scenario. I'm an administrator, I'm trying to update my machine. One of the things that we have done here with Solaris 11 is we've completely streamlined the whole process.
As you go through the different steps of updating the machine doing the planning for that then doing the actual reboot, we've actually gone in, we've completely streamlined that to the point where we can actually do this whole process in a matter of just a couple of minutes, depending on the size of the machine, of course. That may vary, but for a typical size machine, that's something like two minutes, right? That's a huge step up, this means if I'm trying to manage a large-scale environment doing this, having this kind of capability available is absolutely critical. It's helping me to actually minimize the amount of planned downtime that I have to take into account here. The other important fact is that with the way we've constructed this, we've introduced a new packaging system, as I mentioned earlier.
As part of that new packaging system, it tightly keeps track of all the dependencies there. By constructing this on top of ZFS as the default root file system, it allows us to do the process that I just explained where you're going through all the steps. It also allows you that as you go through the steps and something bad happens, your machine goes down, there's a problem with the application for whatever reason, something bad happens, you can just go back, and you just reboot your machine, and you just bring it back up in the previous stage, in the previous state. That means you're never going to end up in a situation where you're completely breaking the system, you have to reinstall it. All it means is that worst-case scenario, you're going to be back in service by doing another reboot.
You're going to be back in your old previous boot environment, the whole cycle takes something like four minutes or so. That's a very critical aspect for customers. Because if I'm trying to-- I know I have a limited amount of planned downtime. I try to upgrade my system, and I want to make sure that if something goes wrong, I can just quickly go back in the previous state without being sort of in a limbo state here. That's something that's absolutely critical. The other important aspect, of course, is doing this in a large-scale fashion. By having a lot of this built into the operating system allows you to quickly automate that, right? Be it with your own scripts, be it by using Oracle Enterprise Manager Ops Center.
It allows you to actually do this now in a large-scale fashion, where you do this across thousands, if not tens or in some cases hundreds of thousands of either physical or virtual machines. You can do that in a very effective way. What are customers actually doing with this today? I mentioned earlier, we shipped Solaris 11 less than a year ago, and we are seeing customers taking full advantage of these capabilities. This is a public services customer. They're running a very mission-critical environment where they have SAP running on top with the Oracle Database on top of Solaris 11 with some of the latest Oracle hardware. They have Solaris clustering across these different nodes for high availability purposes.
One of the other things that we've done since we've shipped Solaris 11 is these other products, Solaris Cluster, Enterprise Ops Center, and Solaris Studio came out. All of these products are now packaged in the new packaging, in the new IPS format. That means if I'm trying to upgrade my Java and there's a fix that's required in the kernel for the new patch level, the packaging system will automatically keep track of that and will enforce that. It will not let me do that update until I've actually applied the new patch level in the kernel. Same thing here. In the case of this customer, they trust the system so much that they're now automatically doing the updates. They have cron jobs running.
They're doing this overnight, rebooting the system, they just come back the next morning and it just shows up in the new state. Those are the kind of innovations that customers are able to do now with their environment, in this case, with a very mission-critical environment, without any of the human intervention that you're used to with other operating systems. This is just a quote from them, patching is so easy that we just automatically trust it, they just update it every week. The other key area that I want to highlight here, of course, in the context of cloud, is having super efficient virtualization there. We have that today with either Oracle Solaris Zones, which is the virtualization built into the operating system, and also the underlying hypervisor-based virtualization that we're doing with logical domains today.
This is kind of a bigger picture here around that. Customers, of course, are used to the high availability, to the zero overhead, the high efficiency that Zones provide them. What we've done now with Update 1, actually, we're actually providing programmable management APIs, which means that you don't have to have a separate management agent running there. If you want to write a script, you don't have to go down the CLI level. We provided that higher-level surface that you can tie into, you can use. It's language independent, so you can use C, you can use Java, you can use Python, you can automatically remotely program your environment and remotely start up Zones to reconfigure them without actually having to log into the Zones console, into the system.
Of course, the other thing is the continued updates, the continued cooperation with the Oracle Enterprise Manager Ops Center team. They are about to come out with their Update 2, which I believe is already out, which is fully integrated with all the latest features in Oracle Solaris 11.1. One of the major enhancements in this context is we've completely streamlined the way you run Zones off of shared storage. We've greatly simplified that. It's very easy now to run Zones, not just on shared storage, but in a lot of cases, customers can now run this in an encrypted fashion, you can just quickly, as you move your Zone over to a new machine, it's very, very easy, dramatically simplified the way we do all the management of that. The other big area, of course, is virtualization of the network.
I'd like to give you a quick outline here of how this has been evolving over time. With Solaris 10, of course, we introduced Zones. It initially came with a shared stack, which means that essentially all the traffic flows together and goes through a shared TCP/IP stack. With Update 4, we actually enhanced that, so that you can actually have dedicated TCP/IP stacks. Everything would still flow together at the NIC level, at the MAC layer. What we then did with 11 is we introduced with the introduction of Project Crossbow and all the network virtualization capabilities, we did two things. One is we actually extended this virtualization all the way down to the MAC layer, we've also pulled in a lot of the networking services into the box itself.
What that means is that with Oracle Solaris 11 today, you can run your environment where you have fully dedicated network interfaces, virtual interfaces. You can have all networking services like load balancing, like routing, and firewalling all built into your host, as opposed to having to have separate physical infrastructure that costs you a lot of money. Then with Update One now, we are extending those capabilities all the way down, all the way out into the network infrastructure. There is a bunch of protocols that we introduced there, which allows us to actually talk to the switches, discover the switches, talk to them, set priorities, then enforce those priorities.
That allows us to do things like convergence of Fibre Channel over Ethernet, but it also gives us a lot more flexibility in the way we can actually prioritize traffic and enforcing SLAs all the way into the network. This was what is happening on the Ethernet side. With the recent acquisition of Xsigo, that also gives us the ability to actually have the latest and the greatest InfiniBand network infrastructure tightly integrated with Oracle Solaris as well as with our storage. What Xsigo allows us to do is that, or allows you to do as a customer, you take advantage of the underlying 40 gigabit bandwidth that you have in the InfiniBand fabric, they will be refreshed at some point to 100 gigabit. On top of that, you can actually emulate your Ethernet traffic and your Fibre Channel traffic.
They have a very, very sleek management solution around that, which allows you to manage that, have it fully integrated with the Oracle Solaris virtual network infrastructure, and you can actually run your whole environment in the most highest performing and also in a fashion where you are reducing the administrative cost as much as you can. It is pretty exciting, I am very excited about this acquisition. A bunch of very smart guys I think that were acquired there, my team is already busy working with them, trying to figure out what is the best way to integrate these pieces together. Storage, I think, is the other interesting piece here, is that it is not just about compute nodes talk to the network infrastructure.
The interesting part here is extending some of these SLAs all the way into the storage infrastructure and providing you the fastest access to your storage box. Whole bunch of enhancements that went into Oracle Solaris 11 on security space. I already talked about immutable zones. Those have been shipping since Oracle Solaris 11 itself. Whole bunch of other enhancements, of course, in 11, with unified crypto engine that allows you to automatically take advantage of the underlying crypto framework. ZFS encryption was another key feature. Whole bunch of different features that you can see on the list here. With Update One now, we are adding a whole range of new features in addition to that. I already talked about the new remote programmable interfaces with what we call a Remote Administration Daemon, that allows you to essentially reduce the number of logins that you need into a machine.
Ideally, the administrator doing the day-to-day maintenance should no longer have to actually get into the machine directly. The other big feature I want to highlight here is the ability to automatically do your compliance checking. What does that mean? A lot of customers that I talk to complain about the increased scrutiny that they're getting today from their auditors. They don't have, in a lot of cases, they don't have very good tools to automate that and to prove to their auditors that everything is fully compliant. With 11 update one, we have a new tool, a new framework that we're implementing there, which allows you to quickly automate that.
You can basically set your policies, based on the policies you set, the new framework will allow you to quickly run a report and automatically come back and tell you whether you're compliant, if you're not compliant, in which areas you have holes, and hopefully convince your auditor, and yourself of course, that you're in full compliance there. In addition to the framework, we have fully integrated that with our auditing. That means that as you're doing these checks, all of this gets pulled out, all of this information gets pulled out and automatically gets audited and provided to you. Another area I want to highlight is CIFS. We talked about some of this last year already with 11. All the major services are there today. With 11, we introduced dedupe functionality, which allows you to compress your data down by orders of magnitude.
In total use cases, we were seeing a 10X deduplication ratio. The other key feature, as I already mentioned, was the data set encryption. With that, of course, we have all the major services in conjunction with CIFS. You have all the file-sharing capabilities, be it through NFS, through CIFS, through various other protocols. All of that is there today. Again, same thing here, whole bunch of different updates that have gone into the 11 update one release. One of the things I just want to quickly highlight here is the thin provisioning for UNMAP, and that means that you can actually dramatically reduce the amount of storage that's used there by quickly freeing up your LUNs there. What are customers doing today?
I quickly alluded to some of, at a high level, what customers are doing, I want to go into a little bit more detail here, and show you what this really means for you as a customer. On this slide here, you see what AAPT is doing. As I mentioned, they're actually providing both infrastructure as a service as well as platform as a service. In this case, on the infrastructure service example, they were actually able, by using the latest zones technology, using all the network virtualization, they're able to achieve consolidation ratios up to 40X compared to a physical environment. 40X compression ratio. That's something that's very hard to achieve with any other virtualization technology, that speaks for the efficiency for the fact that it's the virtualization technology with the lowest overheads out there in the industry.
One of the things that they're also doing, in addition to just achieving these high consolidation ratios, they are also getting additional cost savings by pulling in some of the networking services into the actual physical box. In other words, rather than having a dedicated firewall, physical firewall, dedicated load balancer, this just becomes a virtual function, a virtual environment within your physical box. That's a tremendous cost savings in addition to what you're already getting by the high consolidation ratio. The other example is the example of U.S. Cellular. They have built their infrastructure as a service environment. What they have been able to achieve is they've been using logical domains, and they have typical cloud fashion. They have small, medium, large, 3 different types of VMs that people can check out. They have, again, they fully integrate that with the Oracle Solaris Cluster capabilities.
That means not only are they providing these services in a cloud fashion, but all of these services actually are clustered across, and that gives them a super high level of reliability. They're getting the 10x consolidation, they're getting the 5x performance enhancements that you would expect by consolidating things. But they're also getting this in conjunction with the latest enhancements on the Oracle Solaris Cluster side. This is just in broad brush strokes, an outline of what we have in terms of cloud capabilities. In total, I think we have something like over 300 projects that went into Oracle Solaris 11.1. Again, this is just some of the things that I'm highlighting here, and doesn't really do justice to all the work that went into Oracle Solaris 11.1.
Let me move on and talk a little bit more about what we're actually doing in the context of engineered systems, and also in terms of just a general integration of the Oracle stack. At a high level, if you look at this chart here on the left here, we're continuing to build Oracle Solaris, and we're building all these other components as the best of breed, best in breed components out there. In other words, it doesn't matter whether you're trying to run Oracle app, whether you're running your own in-house app, if you're running a Java app, if you're running a third-party app, we want to make sure that we have Oracle Solaris. Oracle Solaris is the best platform to run any of these components on. The same thing is happening, of course, with all these other products, right?
However, being part of Oracle, it allows us to actually have these kind of deep engineering discussions with our partners, in the Oracle Database team, in the Java team, in any of these other application teams, and have these discussions and figure out how we can actually do engineering that allows us to make the overall experience better, be it in terms of performance, be it in terms of security, be it in terms of availability. You name it, what can we do to enhance the overall experience as opposed to just improving each layer at a time? Of course, in the context of engineered systems, we are actually then taking these components and all these different optimizations across those components, and we're creating an even more restricted environment where you take the fastest connections out to the fastest interconnect with 40 gigabit InfiniBand connections.
We are taking the latest SPARC T4 capabilities of the SPARC T4 processor, taking all these other enhancements in the Oracle Database, and we're building these engineered systems that are tuned and optimized for the best performance and the best security. What do customers actually get out of this? The following slide here just shows you how by providing these engineered systems, how it actually has allowed us to dramatically shrink down the time it takes customers to deploy some of this. The following here shows you that, in some cases, the time it takes to actually install and configure the system has been pushed down to something like less than a week very consistently. These are just different customer examples here. The total time to actually go live, in most cases, is significantly below 8 weeks.
Which for those of you who are used to taking all these individual components and stitching them together is a very impressive result, I think. How are we achieving that? As I said, it's a whole new lifestyle that we have embarked on here. It means that as we are producing, as we are in the design phase of a new product, we have these ongoing discussions with these other partners, and we figure out how can we best optimize the overall solution. Is it something, let's say you have a new performance problem, you're trying to get things, speeding things up to the next order of magnitude of performance to the next higher level. We can actually go in and we can look at, do we solve the problem in the Database? Is it something we want to offload into the operating system?
Is it something that needs to be solved in the RAC layer? Is it something that needs to be pushed down all the way into the physical hardware? Those are the kind of discussions that we're having. I'll give you a lot of examples in the following slides where I show examples of how we've pushed things down into the operating system, how the Oracle Database is taking advantage of some of these features, new features in Solaris, and also talk about how some of these features are actually going to be showing up in the future of SPARC processors also. Of course, as we go through the rest of the life cycle of product development, testing, and then ultimately support, all of these things can be done in concert.
If you as a customer is calling us and you have a problem with your Database performance, chances are we have the same setup somewhere in one of our labs, and we can quickly reproduce that, and we can quickly track down what the issue is. You're dealing with a single vendor, one throat to choke. Greatly simplifies a lot of the traditional problems. This is an eye chart here which shows you all the optimizations or some of the optimizations that we've done so far. A subtitle here says it's the tip of the iceberg. What that means is that it's only been a little over 2 years since we've started working with the Database team. At this point, I have some of my engineers co-located with them, and they're sitting together on a regular basis.
There's a whole bunch of ideas that we're currently working on and that will be showing up in future Solaris 11 releases and ultimately Solaris 12. Next slide. Just a continuation of that. Work in observability, performance, availability, multi-tenancy. There's work going on in all of these different areas in order to provide customers with the best possible experience. Let me give you some examples of what customers are getting out of this. One of the big features that Oracle 12c is going to introduce is the integration with the underlying DTrace capabilities.
What that means is that Oracle 12c, if you're an administrator, you run into an issue, you, an administrator, are now able. Oracle 12c will automatically query the DTrace IO output from Solaris 11, and it will bring that back to the administrator, and they're going to be able to quickly identify, isolate the problem, and figure out whether it's a problem that's in the operating system, it's a problem that's in the database, or whether it's a problem that's outside of the box that, be it storage or be it network. Another key feature, I talked about zones earlier. That's another key feature that we're now enabling with SPARC SuperCluster.
That means you can actually take a SPARC SuperCluster and you can carve it up into secure environments with zones, and it's fully compatible with the 40 gigabit InfiniBand protocols that are running there in order to talk to the back-end storage. That's something that's going to be enabled with SPARC SuperCluster before the end of this year. One other example of this is the integration with auditing. With Oracle Solaris 11.1, we are actually exposing a lot of the auditing information to the next version of the Audit Vault, and with the next release of Audit Vault, that information will automatically be read out and will be fully integrated with the log information of the database. You have a single pane of glass where you can see what's going on in your box across those different products.
As I mentioned earlier, the next level of this is essentially where we're looking at how we can offload some of this functionality down all the way into the physical hardware. By doing that, similar like we had with the crypto, where we actually, by offloading the crypto functionality down into the physical hardware, we were able to speed things up by a factor of four or five. We expect, in some of these cases, we expect similar breakthrough performance enhancement for the Oracle Database by doing that. That's something that's going to be. Obviously, silicon takes longer than software, so this is something that's going to be showing up in some of the future SPARC processors. Without going into too much detail here, similar work is going on with the Java team also. Very excited about that.
One of the key features that I would like to highlight here is the integration of DTrace into the Java Mission Control tool. The Java Mission Control is a way to actually do tracing within your JVM, figure out where the bottlenecks are, and this has been extended now with the release of Java SE 7 Update 4. This has been extended, and you can now have your Solaris plugin and look all the way down into the Solaris kernel and figure out where your bottlenecks are all across the stack. Of course, a lot of the other work going on around performance. This is just an example of Java SE 7, and you see how performance has been improving.
A lot of very dedicated effort here, and we were able to improve overall Java SPECjbb performance by a factor of 2.2 over the course, over the lifetime of Java SE 7. And these are just the two I wanted to highlight here. Of course, we're working very closely with a lot of the other applications teams, figuring out how we can enhance the applications, how we can make them run faster, how Solaris can be the best platform for any of these applications. Following some examples of what a customer is actually getting out of this. I already mentioned Ativas earlier. They were able to get a 10x consolidation ratio, and they were able to actually reduce the overall license consumption by a factor of three. These are not 10% improvements, these are not 20% improvements.
These are some serious enhancements here that we are providing, serious value that we're providing to those customers, and it makes it so compelling to run Oracle Solaris 11 in the kind of a supercluster environment, in conjunction with Oracle Exadata Storage Cells, with all the enhancements from the Oracle Database. It's a very, very attractive offering. The other thing that we're seeing is there's quite a few customers now that are actually moving over to Solaris and to SPARC from some of these other environments. Like KED is an example. They used to run on their environment on an HPE Superdome, and they've now completely moved it over, running that with Oracle Solaris 11, running that in the context of SPARC SuperCluster, and getting, again, 20x consolidation ratios for their business. Significant savings for them, and the list continues. I won't go into all the details.
Another interesting one to highlight is the China Tobacco, which actually went and they did a complete take out of-- they had this existing IBM environment with IBM Power Systems, Db2, WebSphere, and they completely replaced that. They're now running WebLogic and the Oracle Database on top of SPARC SuperCluster, on top of Oracle Solaris 11, on top of the latest SPARC T4 processor. I'm seeing an increased frequency. I'm seeing requests from customers coming to me needing help for how to port over from HP-UX, how to port over from IBM Power Systems and AIX over to Solaris, and we have great tools available. We have porting guides. We have various other white papers. It's a fairly straightforward thing, actually. Here, another example of Oracle. What is Oracle doing internally, Oracle internal IT? This is a great example.
This is the so-called global single instance, which means that's what's processing all of our financial data. That's what's processing all of our HR data. This is processing critical customer data and all that. It's one of these things where if this thing goes down, bad things happen. Larry will be very unhappy, and it really affects a lot of different people. They're in the process of converting that over into a SPARC SuperCluster with Solaris 11. They're using storage expansion racks and also running this in conjunction with Exalogic. They're able to get a huge consolidation ratio. I think they're replacing something like 30 EMC racks and condensing that all into SPARC SuperCluster with the storage expansion racks. For customers, it's a great reference point. It's a great reference story. I'm very excited about that.
I talked a lot about the optimizations that we're doing across the Oracle stack, up and down the stack, and how we're taking some of these enhancements and bundling things up in the form of engineered systems. In the following, I'll be talking about some of the traditional properties that Solaris is providing and how we've enhanced the investment there. One key area, of course, is the area of performance and scalability. There's a lot of features already in Solaris 11 that have allowed us to parallelize the networking stack, providing new I/O capabilities that allow us to optimize things across a very complex piece of hardware. The key feature that we're adding here in update one is the next generation of our virtual memory subsystem.
What that means is that as we are scaling up to these large-scale environments, some of our T4, M4 block-based machines that we're currently running on, some of these machines have 32 terabytes of main memory at this point. That's something where our old VM subsystem just would run into bottlenecks here. With the new next generation VM2 memory management, it actually has new features like the predictor, where we go in, we sample the actual page consumption. Based on the pattern of the application consuming the pages, we can then actually go back, we can pre-assemble some of the pages. That dramatically helps us to speed up things and make sure that we have the right set of pages available, the right mix of pages available at any time.
Some of the enhancements, some of the results we're seeing is a 2x enhancement of our database start and stop time. In terms of the actual memory mapping, the actual effort required in order to do memory mapping, we're seeing a 45x improvement there. Tremendous benefits, and we're confident that with this new functionality, we will be able to scale up to hundreds and ultimately thousands of terabytes of main memory on some of these systems. Just looking, the following slide here is just showing you how over time, how things have grown. With Solaris 9, the max configuration that Solaris 9 would run on was just a little over one terabyte of main memory. 72 threads and 72 cores used to be considered a very, very large box. With Solaris 10, we've dramatically increased that.
We're running on systems now with 4 terabyte of main memory, over 1,000 threads. With Oracle Solaris 11, that allows us to scale up with all the enhancements now in the new VM subsystem. That allows us to scale up to these boxes, to the M4 boxes that are shipping in next year, with 32 terabyte of memory. The next generation after that will have 64 terabyte of main memory. We're doubling the thread count. The number of cores going up to above 1,000. There's a lot of enhancement that we've been putting in, and of course, the road continues here. As we are getting ready for our planning for Oracle Solaris 12, we need to, of course, take into account future systems that are going to exceed those numbers. I gave you an overview of what's in Oracle Solaris 11 today, what is new in update one.
Again, these are just highlights. At a high level, essentially, whole bunch of investments that we've made around cloud, around making it easy to manage clouds, making it secure, making them highly available. I've talked about some of the optimizations that we did with the Oracle Database, with Java, and with some of the other Oracle applications. I've also highlighted some of the investment that's going into some of these traditional properties. All right. With that, I'd like to hand over to my colleague, Bill Nesheim. Bill is going to be talking about the launch of Oracle Solaris Cluster 4.1, and he's going to explain how the new capabilities in 4.1 allowing you to build your highly available enterprise class environment.
Thanks, Markus, for the chance to be here today. Really appreciate the opportunity. My name is Bill Nesheim. I'm the vice president at Oracle responsible for Oracle Solaris platform engineering. Part of my team is responsible for Oracle Solaris Cluster. Today, we'll be talking about Oracle Solaris Cluster 4.1, our latest release of our enterprise high availability suite for Oracle Solaris. Oracle Solaris Cluster 4.1 brings true mission-critical availability to the cloud. We've already heard from Markus how Oracle Solaris 11 is the best Unix for Oracle environments. It's built for cloud infrastructures and is the number one Unix for enterprise applications. Oracle Solaris Cluster is a critical component of the overall portfolio, bringing high availability to the Oracle Solaris environment. Oracle Solaris Cluster has a long history supporting many releases of Oracle Solaris. We released Oracle Solaris Cluster 4.0 last year along with the release of Oracle Solaris 11.
We've released the latest update of Oracle Solaris Cluster, which we've named 4.1, in conjunction with the release of Oracle Solaris 11.1. Oracle Solaris Cluster supports SPARC platforms, X86 platforms, any place you run Oracle Solaris in an Oracle VM environment, together with zones, taking full advantage of Oracle Solaris features such as zones, such as legacy containers, such as any of the kinds of role-based access controls. It's a rich set of availability features for Oracle Solaris 11. Any kind of high availability framework is responsible for monitoring the health of cluster component. What's unique about Oracle Solaris 11 is the rich and deep integration that Oracle Solaris Cluster has with each of these critical Oracle Solaris technologies. Oracle Solaris Cluster monitors the health of the cluster components such as servers, storage, networking, the operating system itself, the network stack, the overall virtual machines, and of course, applications.
This allows the application to tolerate any kind of failure in the environment, be it a storage failure, a network failure, or a complete server failure, or even what's more common and more important, a failure of an application at a given point in a given application stack. Oracle Solaris Cluster has robust algorithms that have been proven over many years, including heart beating, cluster membership, cluster configuration, quorum, device fencing. There's a rich set of capabilities that have been proven in production deployments in Solaris 10 environments. Oracle itself runs our enterprise business applications, our global single instance on Oracle Solaris Cluster. This allows a recovery of the fast recovery of the cluster infrastructure and applications to make sure that your mission-critical environment remains available 24 hours a day. Oracle Solaris Cluster has a rich history of supporting the rich set of virtualization capabilities in Solaris.
Oracle Solaris Cluster supports zones, supports cluster of applications between zones, or even just failing over zones from one system to another. Oracle Solaris Cluster also supports failing over applications from one Oracle VM server domain to another Oracle VM server domain. This provides very efficient availability and multi-tier savings, giving you true multi-tenant application clusters where you can combine application, web, and database tiers into a single environment. This is really critical technology for continuous cloud services. Today, I'm pleased to have with me Meenakshi Kaul-Basu, who is the director of the engineering team responsible for Oracle Solaris Cluster. How are you doing, Meena?
Thank you. I'm doing very well. Thanks a lot.
Really pleased to have you here. We have an exciting set of capabilities that your team has added to Oracle Solaris Cluster over the past year, which are now available in Oracle Solaris Cluster 4.1. Can you comment a little bit on the Oracle Solaris Cluster innovations that we've made?
We've tried to make sure that all the features that we have had in the past, we've actually tried to make that available in Oracle Solaris 11.
Not to forget that there's a lot of innovation in Oracle Solaris 11 already, and we've been trying to make sure that those features and functionality are available in the highly available environment. There's a lot of interesting stuff that has gone into Oracle Solaris Cluster 4.1, which we can talk about if you want.
Yeah
that we've gone ahead and enhanced. There are improved performance enhancements that we've done. We've tried to make sure that applications that can run only in Solaris 10 can be run within the containers inside.
It's a great set of capabilities, but before we start getting into the specifics here, let's talk a little bit about how some customers are using Oracle Solaris Cluster today.
There's a customer that we have in Europe, a public sector customer, that has implemented a multi-tier application environment on the Oracle Solaris environment using Oracle Solaris and zones in that environment. This is a really true multi-tenant installation. Tell us how they're using Cluster.
This particular customer is actually using SAP on Solaris 11. That's something absolutely new. What they've actually gone ahead and done is that they have created zones where they run database in a zone. They have clustered the zones where they run the database on Solaris 11. They also have SAP running in the application tier in a zone again, and they have made that HA. What's interesting in this configuration is that there is a lot of dependencies between SAP and the database, and you want to make sure that the applications come up, and if the database is down, having the application server staying up and running. There are a lot of dependencies that have to be taken care of.
They use Oracle Solaris Cluster to manage the dependencies between each of these applications in a highly available environment.
That's absolutely correct.
That's fantastic.
This is working so well for these folks. One of the unique capabilities of Oracle Solaris 11 is the integrated easy update capability. These capabilities of both Oracle Solaris Cluster and Oracle Solaris 11 are working so well in this environment that these folks actually have automated their regular patching operations.
That's correct.
This is patching not only the operating system, patching Cluster, handling, using things like rolling upgrade capabilities. This is fantastic.
Yeah
To see this. Another example, we have a customer that is implementing clustering across a couple of T4 systems. These are systems that support Oracle VM Server for SPARC. As I understand it, that they're using Oracle Solaris Cluster to provide high availability between these systems.
Correct. It was actually really interesting to hear them at Oracle OpenWorld. They had a session, it triggered a lot of questions from a lot of people who think that they are trendsetters, it's really interesting. These people have done is that they have created Oracle VMs-
different sizes of them. They call them large, medium, and small, they have different applications actually running in them. They have all their applications. They're not just traditional applications, traditional ISV applications. They actually have their own applications as well that they have put into these logical VMs.
That's where the Oracle Solaris Cluster capability comes so much into play, where we can provide availability to even a custom application.
That's correct. I mean, it's pretty straightforward and simple for people to actually create HA, make their applications highly available, and run them on Solaris Cluster.
Okay, this is what people have done with Oracle Solaris Cluster and Solaris 11. Let's talk a little bit about the kinds of enhancements that we've made to Oracle Solaris Cluster 4.1. First of all, one of the major features of Oracle Solaris Cluster is its rich portfolio of application agents. In the Solaris 10 environment, we had a vast array of agents for applications that have been qualified on Solaris 10. I know your team has been working very hard to qualify additional agents for Solaris 11. Can you comment on a few of the specific ones that are now available?
Right. At the time where we shipped the last release, 4.0, there were a certain set of applications that we had HA, and of course, it's most important that the ISV applications, or for that matter, Oracle applications, are available on Solaris 11 Before we can make them highly available.
In this particular release, we've actually almost more than doubled the applications that can be run in this Solaris Cluster, in this Solaris 11 environment. Some of them that I should talk about is we have PeopleSoft Scheduler.
PeopleSoft Process Scheduler.
Correct.
You can make that available so people's batch jobs, you can guarantee that they're going to get finished overnight.
Correct
available in a highly available environment.
PeopleSoft Process Scheduler, as well as the app server is also available. We have SAP that we already.
just talked about. We have the Oracle Web Tier, which is the content manager. That's available. WebLogic Server is available. Again, we have Oracle 11g R2 also available.
Right
supported. I want to point out that we can actually continue to run Solaris 10 applications on Solaris 11 with Solaris Cluster in the branded containers.
Okay, that brings you even a richer set of applications.
Correct
to the Solaris 11 highly available environment.
That's correct.
Can you comment a little bit on the integration with storage? Clustering and storage have had a long history of a very deep working relationship. As people are moving more to a NAS environment, a deeper integration with NAS capabilities becomes even more important. We have a wonderful storage product that provides both NAS and SAN kinds of services, the ZFS Storage Appliance. Can you comment a little bit on what we've done for deeper integration there?
Sure. We had support for ZFS-SA as a storage box, as a NAS box.
Yes.
We have that support already in place for Solaris Cluster, what we've done is extended that to be able to use it with this cluster geographic edition, which allows cluster to be placed in completely different geographies, unlimited distance, and use the ZFS-SA replication technology underneath it. This is a support that we also have in place now.
This is particularly interesting, not just for ZFS customers, for customers who want to use this replication, but also in the SuperCluster context.
SuperCluster, you have Solaris Cluster also part of it to make applications highly available.
Most of these application who need to use storage would use ZFS-SA as their storage box. You can continue to have disaster recovery capability in SuperCluster for such applications using ZFS-SA replication.
Okay
Solaris Cluster is now providing that support.
You have the cluster framework will provide a capability of failing over your entire operations from one data center to another.
Absolutely correct. Yes.
Fantastic. One of the key capabilities of Oracle Solaris 11 is the Zones capability and the ability to do very deep and efficient virtualization. Lots of enhancements went into Oracle Solaris 11.1 around Zones. Can you comment a little bit on the kinds of capabilities that we've added to Oracle Solaris Cluster to take full advantage of the Zones capability in Solaris 11?
Sure. We support Zones as a black box. We support Zones as a zone cluster.
where you can create different zones on different boxes and actually cluster them together and have a whole 3-tier deployment within it. Of course, there are various ways we see customers using it, but the specific enhancements we've done is basically, one, that you can have a zone cluster where you can create a strong branded zone and run
S10 application inside it. We talked about it briefly. Now you can run it in zone clusters as well.
The other major enhancements we've done is support for exclusive IP.
Okay.
Solaris has support for exclusive IP.
Right. Solaris 11 allowed one to deploy basically its own private IP stack for each zone. With a private IP stack, now you can support that in an HA environment.
Correct. That's something pretty exciting.
There are some other enhancements we have done to improve the performance and the management of zone clusters itself, as well as I talked about the Configuration Wizard.
already. That's something that's pretty exciting.
Yeah, that is very exciting.
Certainly, the biggest advantage of Oracle Solaris Cluster in a Zones environment is really there's one installation of Oracle Solaris Cluster. Even in a virtualized environment where you may have many different applications and tiers of your application, there's only one point of management, single point of management, a single point of installation, single point of patching for-
That's-
the HA environment.
That's correct.
That really simplifies the system administrator's job.
Yep.
Other areas that have been a big focus for us in Oracle Solaris Cluster is fast failover.
Making sure that the system can failover very quickly when something bad happens, can you comment a little bit on the kinds of extensions that we've made in that area, particularly around I/O?
Certainly. One of the areas that we've actually made enhancements is if the connection to the disk, for example, to the storage rather, goes down for some reason, we've actually improved the performance significantly, from minutes to seconds. That's pretty significant, where we can detect faster detection of the storage of these outages and be able to make sure that the right thing is done and the failover happens accurately.
Right. You're making the operating system basically behave differently in a clustered environment, where in a clustered environment, the right thing to do is not to retry forever until the disk finally figures out how to give you the data.
That's correct.
The right thing to do is to notify the cluster framework so that an intelligent decision can be made about moving the overall workload.
Yes
from one cluster node to another cluster node.
The faster you can do it, the better it is.
Right. The better the service level.
Right.
It means minimal service interruptions on failover.
That's correct. That's just one of the enhancements we've made there. There are other enhancements we've made in the user land space.
to improve the failover performance of applications. That's something we've gone ahead and done.
It's all about availability.
Absolutely.
It's really important.
Right
if we have the deep knowledge within the cluster framework of what's going on in the operating system, we take full advantage of that to provide better availability for applications.
That's correct. Yes.
Let's talk a little bit about one of the things that this use of exclusive IP stacks and taking full advantage of the Zones enhancements of Solaris has allowed us to do. I understand it, with an exclusive IP stack, you can actually run labeled networking. What does that mean to Cluster?
Right. What we've done is extended that to an extent. Zone Cluster already provides you the isolation-
Right
between the different clusters.
Different applications could be running in each of these zone clusters.
We've done is extended it such that you can be absolutely sure that these are secure environments.
We've used Solaris Trusted-
support in basically making sure that these zone clusters are completely secure.
They're completely isolated, no data transfer between these zones at all, even though they're all being managed by a common high availability framework.
Absolutely. You can have a completely secure zone cluster-
that you just want three people in the whole of the company to be able to access, and you can ensure that.
The different levels of protection that you can ensure.
Yeah. This is particularly important in the data privacy environment that we're all concerned about.
That's correct.
This is a very rich set of capabilities around virtualization, around simplified deployment, around availability and disaster recovery, and with the richest set of agents that are available for any kind of operating system-based cluster environment.
Yes.
Thanks, Meena, and congratulations to your team on this release. Really appreciate you being here.
Thank you very much, Bill.
Okay. Just to sum up, there's real advantages in having an overall red stack of applications, operating system software, hardware, firmware, down to the storage. By putting these components from Oracle together in our shop, we can engineer these components together, we can test these components together, we can integrate these components seamlessly with other Oracle properties such as the Oracle Database, Oracle Real Application Clusters, and Oracle applications. Oracle Solaris Cluster is built specifically on Oracle Solaris, for Oracle Solaris. It is focused entirely on making Oracle Solaris applications highly available. We can take full advantage of deep kernel integration, deep integration into Solaris virtualization capabilities. Oracle Solaris Cluster is tested on a daily basis. With every build of Solaris, we test every build of Oracle Solaris Cluster.
This provides you, our customers, with pre-tested, pre-engineered systems where you can be sure that what you are running is what we are running and what we can support. My thanks to Markus for the opportunity to be here as part of the 11.1 launch. Back to you, Markus.
Thank you very much, Bill. Appreciate that. Thanks for giving us an idea for what's in Oracle Solaris Cluster 4.1. I would like to continue talking some more about, we've talked about what's out there today, and I would like to give you a little bit of an outlook on what we're currently working on. Of course, there's a whole bunch of different things we're working on. I just want to give you a quick update on some of the highlights there. One thing that Solaris has always been very good and famous for has been the availability, just the observability that's built into Solaris. With DTrace, we introduced an amazing tool that allows you to go in and figure out what's going on, the depth of your kernel, figuring out what's going on there.
That's been a hugely popular tool around, both among administrators as well as among the developers. What we're doing here now is we're taking things to the next level, and we're providing the additional focus for the administrator, making it even easier, providing a dashboard that allows administrator to go in and to make some of these capability available at their fingertips, essentially. We're looking at the data that we're collecting. One of the key things we're also looking at is how can we actually process the data over time? How can we shrink it down? What is the right data to collect? We're just going to make it a lot easier to use some of this underlying functionality that's already in Oracle Solaris today. What does this mean?
In the following, is we want to give you an example of the kind of problems that customers can solve with that. This is an example where you're a DBA, somebody calls you, your performance is not good of your database. Some queries are taking too long. With the observability engine, with these new capabilities that we're building now, it will allow you to quickly go in, figure out what's going on, not just in my box, go out to your storage box, figure out whether there's a problem in the storage box. In this case, we get a green light back. It's not a problem in the storage box. You go back in, you're now analyzing what's happening in your I/O, and the request comes back and tells you that there's a problem there. Right?
With that, you actually then go down and you break it down into smaller chunks and figure out where exactly is the problem coming from? How do I do that? The analysis coming back essentially is that the backup process that's running on the machine is consuming way too much of your I/O, and that's what's actually throttling down your application. You go back, you figure out what's the right thing to do in this case. Just put a restriction on the amount of backup traffic that you can have in your box there. Very critical for a lot of customers. One of the key problems I'm hearing from people is that they don't have the right tools today, or the tools they have are just not adequate.
The things we're doing here is making it very easy, giving it to them at the fingertips, and being able to not just look at things within your Oracle Solaris host, but then also going out and understanding what's happening in your storage infrastructure. The other big area that we're investing in is around virtualization, of course, specifically around zones. We're introducing a new archiving mechanism, both archiving as well as templating mechanism, which means that you can quickly archive a zone, you can bring it back up as a virtual image, a virtual machine on top of logical domains, for instance. You can also bring it quickly up again as a physical image, and vice versa. It allows you to be much more flexible with your archiving and also move across, for instance, different hardware architectures like the Sun-4u versus Sun-4v, for instance.
One other big area of investment, one of the things we are focusing on, is actually making zones more flexible and giving customers the ability to have different patch levels across the different zones. That means that some of your zones can actually run FCS, first customer shipment version, others can be running update 1 or later updates. You can patch them independently, that gives you a lot more flexibility there. I talked earlier about the enhancements that we've made on the networking space, how we've actually taken some of the network virtualization, extending that into the network, that's a continued area of investment for us. We are part of the Open Networking Foundation.
We joined them last year, we're working with them, trying to figure out how we can actually make the network virtualization experience much more flexible, how to work with various network equipment with a number of partners that we're working with. Ultimately, what we want to do is actually provide full observability across the network and also provide the SLAs required that allow you to safely talk to your storage box and have dedicated service levels dedicated to you. One other critical area I talked earlier about how Oracle Solaris 11 has dramatically simplified the lifecycle management, how it allows you to do that not just on a single box, but also across a large selection of machines. We're continuing to invest in that. The key investment that we're making here is the ability to actually completely eliminate your downtime.
What this means is that as you're going through your cycle, as you're trying to update your system, you can actually do that while the machine is still running, and essentially you would have zero downtime at all as you're going through the whole system. That would be fully integrated with the new packaging system and fully integrated with the rest of the Oracle Solaris infrastructure. These were, again, just a few highlights of the work that's going on currently on the technology side. There's, of course, a whole bunch of other things going on, as well as a lot of work going on between the different product groups. Last but not least, I'd just like to give you a quick outline in terms of where we are with our future releases of Oracle Solaris 11. We released Oracle Solaris 11 at the end of last year.
We just released update 1, the current plan is to release new updates around an annual cycle. In parallel, we are also doing a lot of planning work around Oracle Solaris 12. We have not pinned down a date for that, we're very busy trying to put that together and come up with a date for that. In parallel, we're also shipping a new update of Oracle Solaris 10, though update 11 is supposed to ship at the end of next year, and that will be aligned with the release of the future hardware of the M4 and of the T5 systems. I hope I was able to give you a good high-level overview of what features, what functionality is in Oracle Solaris 11 today, as well as 11 update 1. Please stay tuned.
We're going to have some follow-on sessions where we're going to go into some more detail about what's in Solaris 11.1, as well as in Solaris Cluster 4.1, so please stay tuned for that.