Guest Post from EMC: Microsoft SQL Server 2012 & EMC VMAXe; A Perfect Match
Successful companies do their best to stand out, to provide offerings of value and to balance quality, functionality and cost. Of course, as individual consumers, we understand that “new and improved” is no guarantee of practical, useful, simple and effective. So major product announcements keep most of us a bit skeptical ‘til we have a closer look.
Sometimes we’ll be disappointed.
Sometimes we may be both surprised and delighted.
Microsoft is releasing SQL Server 2012 with an emphasis on important customer needs: “Mission Critical Confidence”, “Breakthrough Insight” and “Cloud on Your Terms”. All of this with flexibility and at scale.
These messages respond to an accelerated business reality. Respond now to new demands. React now to threats – business or technical. Prevent security breaches. Be prepared for continuous operation, even if disaster strikes. And do it all while managing exponential information growth.
As customers create these critical business environments, Microsoft and EMC combine to supply the critical technical necessities. And there are many.
Here are a few examples:
- High Availability and Disaster Recovery with support for Microsoft AlwaysOn Availability Groups and Failover Clustering support across multi-sites and multi-subnets with storage replication
- Backup & Reporting that integrates Volume Shadow Copy Service (VSS) and Virtual Backup Device Interface (VDI) with clones for fast backup, instant restore and Read/Write copies for reporting, testing and surgical restores
- Five 9s stability through rigorous EMC testing that complies with the SQL Server I/O reliability program
- Maximized performance through SQL Server 2012 I/O characterization and array testing that continuously helps to improve storage algorithms
EMC calls this a journey of transformation and we all know it takes some time and some fortitude. But, while it isn’t a straight line, EMC and Microsoft are combining in ways that make the outcome more certain. Fewer twists and turns.
SQL Server 2012 and VMAXe are proof.
SQL Server emphasizes cost efficiency, reliability, availability, simplified deployment, security, predictable performance and the capability to grow, especially in cloud environments.
So does VMAXe.
VMAXe combines simple self-tuning in an advanced enterprise class storage system at a very cost-effective price.
The Microsoft virtual launch event happens on March 7th but there are already customers with SQL Server 2012 deployed. Browse their case studies and you begin to get a feel for the real business impacts in a “mission critical” environment:
We “must access and share operational data with global employees and partners all day, every day” with excellent service and no interruptions.
We need “to support ambitious growth plans, … increase profitability, simplify compliance, and gain better insight into customer relationships.”
We need “to increase availability for ... mission-critical online event management applications.”
These are business environments where saving just twenty minutes of downtime in a year “is huge for us and our customers.”
Always on
So, first order of business is absolute reliability. SQL 2012 is designed for mission critical application environments. VMAXe has a Symmetrix pedigree with all the reliability, availability, serviceability and performance of that bloodline. Its’ multi-controller architecture provides for scale out in capacity and up in performance with five nines of availability. This system architecture means that even with a rare component failure, the system continues to function with little or no performance impact.
Because of their critical business importance, these application environments are typically configured for high availability and disaster recovery. That means a redundant design with the capability for fail-over to a second or third site. Together, SQL and VMAXe have all the fail-safe functionality to provide that peace-of-mind.
Previous SQL environments may not have required this level of horsepower and uptime. Becoming mission critical, now they will.
These are also Big Data environments and SQL 2012 is designed to create the “breakthrough insight” optimized for Big Data applications. VMAXe compliments that with all the muscle Big Data needs to gain that insight quickly, without impacting normal operations.
Virtualized
There’s more, especially for cloud and virtual environments. A feature called FAST VP in VMAXe smoothly automates storage tiering to assure optimal performance for critical workloads. It places the most important data on high performance Flash drives and pushes aged data to lower, less expensive tiers but keeps it within easy reach. It also simplifies array management and reduces weight/space requirements and saves energy and reduces total cost of ownership.
Cloud on your terms.
VMAXe has features that make SQL layouts simpler, backups more flexible, and overall storage use more efficient when multiple data copies are required.
And while some SQL environments may be large enough to require an entire VMAXe, many customers will want to run other applications on the system simultaneously.
There’s a lot of flexibility here to do that.
Storage groups can be set by application priority.
SQL can have priority access to Flash storage.
Cache management and backend IOs can also be prioritized by app.
Need more storage capacity? Add more drives, non-disruptively.
Need more horsepower? Add another engine, non-disruptively.
Your particular use of SQL may be new to mission-critical requirements. Microsoft and EMC are not. We have been doing this for a long time, decades.
So go ahead. Get on the cloud. Deploy for mission critical. Embrace your Big Data. We can help you make the transformation secure and comfortable. Maybe not a straight line, but a well-traveled path nonetheless.
Engage with our experts at: https://community.emc.com/community/connect/everything_microsoft?view=overview
Dick Sullivan
Director, Partner Enablement
EMC Enterprise Storage Division
Blog: http://energymatters.typepad.com/cloud_matters/
Measuring the IO Capabilities of the HP Enterprise Database Consolidation Appliance
I am Ravi Ramachandran from the SQL Server Appliance Engineering team and would like to show how to measure the IO capabilities of the HP Enterprise Database Consolidation Appliance which is optimized for SQL Server.
The Engineering teams from HP and Microsoft collaborated closely for many years to design the DBC Appliance from the ground-up. Many database implementations and customer workloads were analyzed in-depth in addition to collaborating with the SQLCAT (SQL Customer Advisory Team) and select TAP (Technology Assurance Program) partners to arrive at a very robust, scalable and fault-tolerant Private Cloud Appliance, which is operational right out of the factory. The Appliance is the best in breed capable of handling a variety of demanding SQL and customer workloads.
The Appliance ships as half rack, full rack and multi rack. The full rack configuration has 192 Processor Cores, 2 TB RAM, 396 Disk Drives, 57TB Raw disk space, 24TB formatted disk space, and can produce 60,000 Random 8K IOPS at 60% Read, 40% Write. The half rack, which is exactly half of the full rack configuration, can produce 30,000 Random 8KB IOPS at 60% Read, 40% Write. The Appliance can scale up to 10 full racks in half rack increments yielding the multi rack configuration.
The Appliance has a System Center management stack that facilitates P2V and V2V migrations, new VM provisioning, monitoring and alerting, besides providing other management capabilities. Each System Center component, SCVMM, SCOM, SSP, etc., is installed in its own VM and is resident on the Appliance. This management stack overhead is very minimal thereby providing all the compute power and storage capacity available to customer workload VMs.
In our internal testing, we have driven many complex workloads on the Appliance. In addition to evaluating the IO capabilities, around 865 virtual machines running concurrent SQL Server 2008 R2 workloads were pushed on to the full rack to stress test the sustained IO and performance capabilities. This exercise was repeated with SQL 2012 (pre-release bits) as well, to ensure that the performance of the Appliance was in-line with the design expectations.
As part of measuring the IO capability, many publicly available tools to drive those workloads were evaluated. One such tool is IOMeter, which provides the ability to drive mixed Read/Write workloads and is freely available for download from the web. Anyone can point the IOMeter to a functional Appliance (half, full, or multi rack) and run a few basic tests to prove to themselves the Appliance’s real IO capabilities. The step-by-step procedure for set up and how to benchmark the IO performance of all rack configurations is documented and available.
The Appliance IO capabilities are built into the hardware from the ground up, and performance numbers mentioned previously can be obtained with all the System Center management VMs operational. The video is a practical demonstration of running IOMeter against a stood up full rack. You can see the actual IOPS produced by the full rack as well as the IOPS data shown for the half rack.
More information on the HP Enterprise Consolidation Appliance can be found here.
Ravi Ramachandran
Senior Software Design Engineer in Test
SQL Server Appliance Engineering Team
