By MarkusFlierl-Oracle on Sep 25, 2013
It all started on Sunday with a customer advisory board in the San Francisco office. Some of my technical leads and I spent 1/2 day with key customers reviewing our key investments and getting their feedback on those. One of the customers gave a talk on their self-provisioning IaaS and PaaS cloud taking advantage of the latest S11 features. They are using a combination of Solaris zones as well as OVM SPARC in order to achieve super-high compression rates and drive the cost/VM well below what they can get with RHEL and VMWare on x86 hardware. Not to mention all of the other benefits of Solaris like higher up-time, security, super-efficient patching etc.
Later that day I went to Larry Ellison's Opening Keynote. Larry did a great job highlighting how the new In-Memory capabilities of Oracle DB 12c will be taking advantage of the 32TB of DRAM that the newly released M6-32 has to offer. 3,072 threads on 384 cores! When we initially tested the Oracle DB on this large machine it took 2800s to fill a 30TB SGA when running Oracle DB 11g on S11.1. Working closely with the Oracle DB team we have been able to get this down to about 130s now! That's a 21x speed-up! We had to do a whole bunch of work on both sides to accomplish that!
During my general session on Monday I started out by highlighting the accelerated pace of product development at Oracle reflected by the long list of new product releases since last year's OOW: Solaris 11.1 in October, the T5/M5-based systems in March, the refresh of SuperCluster to T5-8, the release of ZS3 (aka ZFSSA) in early September, the release of the M6, the M6 SuperCluster as well as the T5-8 Exalytics at OOW - what a line-up of new products in just 12 months! A great testament to the increased investment that Oracle is putting into the former Sun products! I continued to talk about the increased customer momentum behind Solaris 11 and our latest SPARC hardware. We now have over 4,000 customers running on Solaris 11. Allied Irish Bank is a great example: They consolidated a bunch of older Oracle databases running on older HP Itanium as well as IBM Power boxes onto Oracle DB 11gR2 on Solaris 11 and SPARC. This has allowed them to not only achieve consolidation rations of 10:1, but also provision 10x faster now. We also have a number of good examples of customers moving their RHEL/ VMWare/ x86 estate over to Solaris 11 and the latest SPARC hardware, increasing customer response times by 50% and getting 2x the performance out of the same number of Oracle licenses. Aside from external customers we also have Oracle IT as a major customer: Oracle currently runs its Global Single Instance of E-Business Suite on a SuperCluster taking advantage of all of the optimizations between the Oracle DB and Solaris as well as SPARC. So far they haven't seen any outages at all (I'm keeping my fingers crossed so I won't get woken up at 2 in the morning :-) ). Oracle IT is in the planning to switch over to the latest M6 SuperCluster over the next few months.
Then Mike Montemorano from Verizon Wireless joined me. He did an excellent job explaining why Solaris is strategic for Verizon Wireless: Mike explained how Verizon Wireless has grown their number of physical SPARC servers by almost 50% over the past 4 years and the number of Solaris VMs by 5x. They recently deployed 30 T5-8s and 15 T5-4s to run their mission-critical IaaS cloud. Aside from pure economics he said that their mean time to resolution (MTTR) is 4-8x faster on Solaris compared to other platforms.
I continued to talk about Solaris futures: Over the past couple of years we have greatly expanded the way we think about Solaris. We no longer look at it as just a mere OS, we now look at it as a comprehensive cloud platform: OS + OS Virtualization + SDN + Storage Virtualization plus Cloud Management tools, configuration, compliance reporting etc. all in one! It doesn't make sense to compare Solaris against RHEL, that's like comparing a spare part vs. a full solution. Solaris is the ideal platform for any kind of SaaS, PaaS (Oracle DB, Oracle Middleware, Oracle Java) as well as IaaS.
Here are some of the highlights of what's coming in the next release:
1. kernel zones: With kernel zones customers will have to option to run different kernel patch levels across different zones while maintaining the simplicity of zones management. We'll also be able to do live migration of kernel zones. All of that across HW platforms, i.e. kernel zones will be available on both SPARC as well as x86. Key benefits of kernel zones:
x Low overhead (Lots of optimizations because we run Solaris on Solaris)
x Unique security features: Ability to make them immutable by locking down the root file system
x Integration with the Solaris resource management capabilities: CPU, memory, I/O, networking
x Fully compatible with OVM SPARC as well as native and S10 branded zones
x Comprehensive SDN capabilities: Distributed Virtual Switch and VxLAN capabilities
2. Unified Template Builder: This will allow customers to go from any-to-any of the following: Bare metal, OVM, kernel zone, native zone. For instance: You'll be able to take a zones image and re-deploy it as a bare metal image, kernel zone or ldom. Or vice versa! Pretty powerful, huh? Unified templates also provide us with a great foundation to distribute key Oracle applications as a shrink-wrapped, pre-installed, pre-tuned an configured image where customers can specify at install time whether to turn them into a bare metal image, a zone, a kernel zone or an OVM.
3. OpenStack and Puppet: We are planning on integrating both of these into Solaris. This will allow customers to seamlessly integrate a Solaris/kernel zones or zones or OVM environment with other OpenStack platforms.
I concluded by highlighting some of the differentiating features we are now available in Oracle DB 12c when running on Solaris 11.1:
x Zero-downtime DB re-sizing
x Full-stack analytics: Running on Solaris allows DBAs to quickly identify I/O outliers. For more details check out:
https://blogs.oracle.com/solaris/entry/database_12c_and_solaris_dtrace x RAC lock off-load into Solaris kernel: This reduces latency by about 30-40%.
Not to mention the 100s of other features and improvements that I won't even mention here.
Solaris is on a mission to kick butt! More to come...