HP Storage: Delivering Storage without Boundaries

Editor's Notes

• #2: SWD Key message at Discover 2011: “Know No Boundaries”
• #3: HP Storage solutions within a proven, winning Enterprise Cloud System Change from IT-driven to business-driven storage planning SAN Virtualization made scalable and simple with HP Storage P4000 IP Storage X9000 Network Storage P6000 EVA Storage 3PAR Utility StorageChoose the right platform – P4000, X9000, EVA, or 3PAR
• #4: Let’s examine some main storage challenges that need to be addressed by today’s IT administrators.First, we are dealing with more data than every before, and it’s not getting any easier to manage. In particular, we can look to
• #5: Notes to presenters: Most likely, customers have already heard HP’s ‘Converged Infrastructure’ story and may already be familiar with the concepts on this slide. The idea here is to tie this slide more closely into storage. Key questions:Where do you want your IT dollars to go? Operations (70% of most IT budgets) or Innovation (30%)? What is preventing you from achieving this now? Possible answers: IT Sprawl. Legacy architectures. Inflexible IT processes. Key point: Storage is a key driver of these issues. Digital storage is doubling every 18 months. And storage has a huge ‘operational-cost’ side. A lot of technologies have promised to address this 70/30 ratio but have had little impact. Particularly in storage. To address this fundamental problem there needs to be a fundamental change in storage arc The chart on the slide shows the amount of money spent on various storage operational costs for every dollar spent on the storage hw/sw itself. Note: The source of the data in the table is IDC.
• #6: Customer tell us storage it too complicated, inflexible, inefficient, an disconnected. By bringing together our HP intellectual property and our recent acquisitions, we’re creating storage that’s:SIMPLEfast and flexible response to business needsAGILEfor unpredictable and multi-tenant use casesEFFICIENTfor optimal use of available storage resourcesINTEGRATEDoptimization across rest of IT infrastructure
• #7: P4000LG CNS – Korea (VDI) Q2/3Sales Team: ObjectiveTo provide a virtual desktop infrastructure to serve 8000 8000 users (800TB  of high performance capacity )Initial Customer dilemma:In FY09 Q4, LG CNS execute their internal VDI project with HP Blade server & EMC NS960 storage, however  performance degradation issue on their EMC NS960 was of major concern to themApproach LG CNS gave HP a chance to test HP P4300 to LG CNS, which allowed the SWD team to re-engage in this project and demonstrate the performance of the P4000 HP StorageWorks G2 P4500 x 112 units racked HP C7000 x 28ea - (BL460c x 16) x 28eaStorage Value: $US500k The Bottom Line Substantial improvement in performance as when compared to EMC NS960, the P4000 G2 proves much better IOPS on OS boot testing and antivirus scan and excellent response rates 
• #9: So what we see happening is that the world of technology is changing. It's a huge trend. The pendulum is shifting to a new paradigm and new strategic opportunity. We’ve seen these periods of divergence/convergence beforeThere is a lot of noise in the marketplace…and lots of different directions But as customers standardize on fewer, strategic vendors, they won’t take risks. They will require proven innovation/results in real-world environments. That’s why HP will win, because we’ve been on this path…for years. Our path for customers isn’t a do over or a blast from the pastInstead of having individual stacks, like we've had for a long period of time, a storage stack, a server stack, a networking stack, these are all now going to be reunited. So what you'll start to see are solutions that come out, that integrate storage, networking, servers, management software, even facility power and cooling.When we talk power and cooling, or virtualization or management, it's not just for the product itself, but for the entire data center. We call our strategy, the HP converged infrastructure. And we wholeheartedly believe this is the way the data centers of the future are going to get built.The HP Converged Infrastructure is a next-generation IT architecture that:Converges virtualized compute, storage and networks with facilities into a single environment optimized for any workloadInside, all resources and processes are controlled by a shared-services engine that provisions and adapts application environments to instantly respond to business demands This architecture helps businesses accelerate the delivery of application environments in a predictable, repeatable way that makes the most efficient use of IT, facility and staff resources to drive business innovation. We believe this strategy will help business unleash their full to potential to . . . Deliver any application, anywhere, on the flyFlex resources on demand in an optimized wayUnleash productivity of administrators, systems and facilitiesProvide predictable, continuity of serviceAccelerate time to business value from IT investmentsSimplifying the data center requires the IT stacks to be brought together into pools of virtualized resources s that are shared by many applications and services. The solution to sprawl lies in a converged infrastructure that reunifies business, application, and infrastructure functions, delivering faster time to application value, simplified management, increased utilization, and lower power cost across an extensible scalable set of applications. This convergence isn’t just about the virtualization of servers, storage systems, and network connections. Convergence also requires bringing together management tools and processes so resources can be managed in a holistic, integrated manner. In addition, to tackle the growing complexity and cost of energy, the next generation data center needs to synergistically link power, cooling and facilities with IT.TRANSITION: This is not something that's going to happen five years from now, or ten years from now. It's actually already happening today. You just might not realize it yet. HP has been leading the trend for more than 5 years, delivering the building blocks of a converged architecture, like HP BladeSystem, Virtual Connect Mission Critical Superdome, virtualized storage, shared-service management with BladeSystem Matrix. HP is the only vendor with the intellectual property and investments across servers, storage, networking, power and cooling and management to do convergence right.
• #15: To better understand the benefits of the LeftHand SAN we need to look at the five key features include with every system. Storage Clustering for scalable performance and capacityNetwork RAID provides your storage with beyond the box availabilityThin Provisioning to help reduce costs by increasing capacity utilizationSnapshot for advanced data protection and a reduction in tape storageRemote Copy for disaster recoveryWe’ll start on the next slide by explaining the benefits of storage clustering.
• #17: ServersUp to 8 blades servers per enclosureHypervisorsD2DTapeStorage2 x P4000sb storage bladesConnected to 70 drive shelfNetworking2 x HP Virtual Connect Flex-110Gb Connectivity
• #18: When it comes to data protection, traditional methods such as HW RAID and redundant components are not enough. Hardware RAID and redundant controllers wont protect your data from human-error, power or cooling issues. Your data must be protected beyond the box, in the event that entire node goes off-line. P4000Network RAID allows you to synchronously replicate your data across the nodes in a cluster for high availability. This protection is granular, allowing you to assign it on a per-volume basis. There are several Network RAID protection schemes that you can choose from based on a volume's performance characteristics availability needs. The first, and most common type is Network RAID 10. In this scheme, the data in a single volume is mirrored across the cluster either 1, 2 or 3 times. The replication occurs real-time, and protects the data from any single point of failure up to, and including, single or multiple storage nodes. Network RAID 10 works well from a performance perspective with mix workloads.The second type of network RAID is parity based. In this scheme consists of a stripe and parity layout that allows you to start with a few as 3 nodes. This type RAID works best with highly read intensive workloads, like file shares. There is also a dual-parity version called NR6.You can also stripe the data with no protection for chosen volumes.It is very flexible by allowing you to change your NR type on the fly when you choose.
• #19: Volumes need to show volumes the same on the nodes from previous slides, show 2mms latency on connectivity, show NW10. Be clearer that the site had failure, make the VM movement more visible, horizontally (done) P4000 SANs store logical volumes in a way that makes HA solutions low cost, low risk, and automatic. P4000 SANs don’t replicate volumes. They replicate the data within each volume. The result is that the same logical volume is available before, during, and after a failure, with failover and failback both automatic and completely transparent to application servers. This makes P4000 SANs able to support virtualized environments not just through the failure of an entire array, but also through the loss of an entire data center. All of this is accomplished by configuring the SAN to be split between racks, rooms, floors, buildings, or sites.P4000 SANs implement fault tolerance using a combination of storage clustering and Network RAID:Storage clustering creates a scalable storage pool by aggregating the resources of a number of storage systems into a single storage cluster. The cluster accepts and responds to iSCSI requests as a single unit. All physical capacity and performance is aggregated and is available to the volumes created on the SAN. When more storage is needed, additional storage systems can be added to the cluster, and the cluster seamlessly, non-disruptively reorganizes its storage to incorporate the new system.P4000 SANs store logical volumes in a way that makes HA solutions low cost, low risk, and automatic. P4000 SANs don’t replicate volumes. They replicate the data within each volume. The result is that the same logical volume is available before, during, and after a failure, with failover and failback both automatic and completely transparent to application servers. This makes P4000 SANs able to support virtualized environments not just through the failure of an entire array, but also through the loss of an entire data center. All of this is accomplished by configuring the SAN to be split between racks, rooms, floors, buildings, or sites.Network RAID synchronously replicates the blocks that make up a logical volume across one or more storage systems in the cluster for high availability. A Network RAID level of 2 causes every block that makes up a logical volume to be stored on two storage systems. Now a single component can fail, a storage system can fail, or in the case of a geographically split SAN, an entire location can fail, and each volume’s data will remain available. Failover and failback is automatic and transparent to application servers because the cluster itself keeps track of which storage blocks are up to date, and it automatically updates outdated blocks when the failed part of the cluster is restored.
• #23: We need an additional slide to show VSA connected to existing FC, DAS or iSCSI storage, and break this down to VSA High Availability Remote site and then showing StoreOnce as DR for Remote sites, non-HA add StoreOnce target into viewShow different types of remote sites as DR or smaller vs. HA sites coming back to a StoreOnce ArrayRemote Copies are based on snapshots. A Remote Copy is essentially a snapshot that is copied to a remote cluster. Because they’re based on snapshots, they also take advantage of thin provisioning and do not require reserve space.If you have a primary site with volumes and snapshots and a second DR site, you can copy the volume and snaps to the remote site via your WAN linkRemote copies can be scheduled based on your DR needs. The initial data copy copies the entire volume – Each additional remote copy sends only the changed data to the remote site.Remote Copies can be run over most any WAN link, and can take advantage of WAN encryption (VPN) and/or compression (Riverbed)Bandwidth can be managed through the CMC by editing the management group properties.Like snapshots, no space needs to be reserved ahead of time on the remote clusterYou can mount these remote copies as fully readable and writable volumes, so that if the primary site goes down, you simply attach the remote copy to a host and continueA wizard guides you through both the failover and failback scenarios
• #26: Slide Purpose: Setup the value prop for the next three slides
• #27: Slide Purpose: explain how 3Par is an efficient arrayThin: With ThP, ThP conversion and ThP persistence and you can but less storage up from convert full LUN to thin and keep LUNs thin over time. Saving space and powerOptimized: With Dynamic and adaptive optimization, you can automatically move data to the correct tier of storage. This optimize the use of high performance drives (SSD) and lower cost, lower power and higher capacity SATA drives.Green: By using 3PAR Thin technologies and data tiering, IT can buy less storage and use it in an optimum fashsion to save power and reduce costs.
• #28: Slide Purpose: explain the simple management Autonomic is the bodies way of reacting to increase load. As you walk up a flight of stairs, your breath deeper and your heart pumps faster. You don’t have to think about it, your body just does it .With 3PARs autonomic management, you don’t have to think about the little details as the array will do that automatically. Setting up a LUN in as little as 15 seconds, begin able to deliver high performance under all applications and failure scenarios and being able to adapt quickly to change. That is alll designed into the HP 3PAR arrays from the start.
• #29: Slide Purpose: Explain the scaling and virtual private arrays featureWith multiple nodes and our unique ASCI, which I will talk about in a minute, the HP 3PAR arrays can scale to meet the demanding and different types of workloads across multiple business units. With the Virtual Private Array feature, workload can be logical isolated for security across business units , yet still gain the benefits of the massive parallelism the HP 3PAR arrays are designed to deliverResilient, requirements change and being able to quickly and easily adapt is key when multplie customers or business units share and array. With Adaptive and Dynamic Optimization, the array is ready to handle those changes without disruption to the business.
• #30: Slide Purpose: Explain the 3 key factors that make 3PAR unquieI spoke a minute ago about the Automomic management and that is common acoss all HP 3PAR arrays.What might be the single most important piece of hardware in the arrays is out third generation ASIC. Every node has an ASCI built in. This ASIC provide many functions within the array. First, the ASIC provides high speed node to node communication. This enables the parallelism of workloads across all nodes. Next is the RAID calculations are done in the ASIC ar wire speed enabling very high performance even when using RIAD 5 or 6.And lastly, Thin is built in. Our ASIC has zero detect as part of the design that allows us to do thick to thin conversion of LUNs at a very high rate. We will talk more about this in the next couple of sldies
• #33: Slide Purpose: Overview the product familyThere are 4 model in the current 3PAR array family. All use the same firmware, software and management. Starting at the F200 with 2 controller nodes and scaling to the T800 with 8 controller nodes. There is an array that can meet you currnet needs ansscal as you grow.
• #34: Slide Purpose: This slide shows the Horizontal and Veritcal applications where the HP-3PAR technology fits best.
• #35: This is your wrap up slide. I suggest that we finish by stating again that our technology has been built from the ground up to support Thin Provisioning which allows us to do more data reduction then other technologies on the market.
• #37: Notes to presenters:The Enterprise Storage Group is located near Boston and is one of the most respected storage analyst practices in the industry. Their message is clear: Unstructured data is growing and virtually all industries and companies are affected. IT budgets are being squeezed.A new model is required. See chart: 2009 was the cross-over year where more unstructured data was digitally stored than structured. The chart shows the relative amounts of structured and unstructured data from 2005 to 2012.
• #38: Notes to presenters: This slide covers 4 key unstructured-data requirements that typical tier-1 storage would have a difficult time meeting. Key points:Customers can put their unstructured data on the legacy, tier-1 storage, but it will end up costing them a fortune in operational costs and most likely still not meet their needs. Mainly because of its massive growth and size of the storage pools, unstructured data has its own special requirements. Downtime as a result of constant storage expansion is just one example:ApplicationCost of downtime per minute: Trading (securities) - $73,000: ERP - $14,800E-commerce - $12,600Supply chain - $11,500Point of Sale - $4,700Email - $1,9004. Massive scalability causes problems for customers because of the space, power, and sys admin support required. Plus, all these spinning disks do ultimately represent a lot of different clients’ data, and thereby a lot of potentially dissatisfied customers if things are not administered efficiently. There are several companies that sell very large storage systems. However, these systems will all top out performance long before their capacity limits are reached. Since, with unstructured data we are talking about potentially petabytes of storage, it is important that there be no upper limits to the performance either. To some extent all four of these ‘needs’ have a cost factor to them and ‘management of huge data sets’ is no exception. How do people manage a petabyte of storage? Do they have to manage multiple, discreet storage systems or can the storage all be managed as one? Can they move the data from tier 1 disks to tier 2 or 3? Can they load balance across all the systems so that performance stays consistent? Can they make copies of the data for protection from data loss? With huge data sets, all of these are serious challenges. Finally, cost containment is an ever-present issue that, because of the sheer size of some of these deployments, has forced everyone to look at space, power, administration costs, and long-term cost-of-ownership numbers as closely as purchase price.
• #39: Presenter notes: Key points.Unstructured data permeates virtually all industries. Only Scale-out NAS can satisfy the high-capacity and high-performance requirements of unstructured storage while controlling costs. These requirements span many applications and many industries. Some of these applications require easily expandable performance. If so, scale-out NAS is the right answer. Some of these applications require high capacity. And some require both. In all cases, scale-out NAS offers advantages over traditional scale-up solutions for many environments. (Depending on the customer you are presenting to, you might choose to emphasize one of the industries on this slide).In the next several slides we will be looking at the specific advantages offered by HP’s scale-out X9000 NAS solution. The following background information is from:From the IDC research document titled: “IDC's Enterprise Disk Storage Consumption Model: Analytics and Content Depots Provide a New Perspective on the Future of Storage Solutions”Transactional-type applications such as ERP, CRM, and OLTP that typically create and access structured data still remain majorconsumers of new storage capacity; however, new applications that generate and access unstructured data are emerging as much morevoracious consumers of storage capacity in typical enterprises. These include:● Digital imaging and archiving systems in industry sectors such as healthcare● Collaborative applications such as email, SharePoint, and product life-cycle management● Video-intensive applications such as video surveillance In addition, recent years have seen the appearance of neworganizations whose business is almost entirely based on the storage, organization, long-term retention, and timely retrieval of unstructureddata. These include:● Companies that engage in high-performance or technical computing such as oil and gas, life sciences, entertainment (e.g.,rendering), government, and other simulation applications● Internet, telecom, and hosting companies that provide large content depots for online audio, photo, and video repositories; email;online backup; online ediscovery analysis/archiving; and other storage-as-a-service offerings for consumers and businessesThe continued expansion in creation and archiving of unstructured information is already having a dramatic impact on organizations'storage environments, with IT executives coming to the realization that managing such file-based, unstructured data will become the primary(©2008 Industry Insights, an IDC Company #214066 Page 5) task for many storage administrators in datacenters. Theseadministrators need storage solutions that can: ● Support faster access to significantly more and larger files for much longer periods of time● Ensure that the sharing of business content is more cost-efficient and secure● Enhance (aka monetize) the value of unstructured data through more sophisticated business intelligence and analytics solutions(IDC' s E n t e r p r i s e D i s k S t o r a g e C o n sump t i o n Mode l : S e t t i n g B o u n d a r i e s a n d E x p e c t a t i o n s)All of the challenges noted above already have a number of significant effects on the storage industry:● They are driving continued large expansions in the amount of storage capacity shipped (the number of terabytes shipped in 2007was 58.3% greater than in 2006).● They are spurring a proliferation of new storage systems, a wave of new storage technologies, and a number of new storage systems suppliers.● They are forcing IT executives to rethink many storage investment and management practices.
• #41: Note to presenters. This is a build slide. Be sure to study it before presenting. If you understand the X9000 scale-out NAS functionality, you can spend a lot of time on this slide. Each click highlights another feature of HP’s scale-out technology. This is where you describe the advantages of HP’s scale-out vs. Scale-up technology. Here are the key points you should cover:The X9000 offers a single namespace where all the storage in the cluster looks and manages like one big pool of storage. You can keep adding even different size systems to the cluster without downtime. The X9720 can grow to 1.3PB, and you can add multiple X9720 systems to the same cluster.You can also add X9320 capacity and performance systems to the same cluster.And, if your environment includes an FC or iSCSI SAN, you can front-end those storage systems with a X9300 gateway and wrap all that storage into the cluster, as well. And eventually expand the cluster up to 16PB. And also allow the entire cluster to then take advantage of the SAN backup to tape.
• #47: The X9000 series is for enterprise customers with fast growing unstructured data who suffer from performance bottlenecks, varied workloads and require scalable capacity via a single mount point.The HP StorageWorks X9000 Network Storage Systems scale out beyond traditional network attached storage in both capacity and performance. Designed to be scalable, flexible, and cost efficient, HP X9000 Network Storage Systems deliver excellent performance and a modular storage infrastructure to accommodate unprecedented storage growth and performance. Central management allows the administrator to manage not just terabytes but multi-petabyte environments at the highest density possible from HP. The pay-as-you-grow architecture and an all-inclusive feature set reduces storage over-provisioning or expensive add-ons. X9000 is the platform for customers designing the next generation data center.Key Features:Data tiering: Migrate files from one storage tier to another within the same global namespace based on file characteristics and attributes. This saves costs by moving seldom-used data from higher cost storage to more affordable storage, thus lowering operational costs and improving overall storage efficiency.Continuous remote replication: Site-to-Site replication to provide continuous parallel data replication from the primary to multiple remotes sites. Ensures quick recovery and high data availability, even in the event of a site failure or disasterRebalancing: Dynamic, non-disruptive data redistribution across physical storage elements to provide balanced storage utilization. Balances IOs across nodes and improves performance by avoiding hot spots on existing storage and when new storage capacity is added.Easy scaling: Add performance and capacity appliance online, with no downtime. Control costs by growing the system according to your needs. Keep storage within a single namespace for easy movement and management.Management— While most large storage systems require multiple management tools to manage the system, the X9000 management server enables a single point for system configuration and control, which includes monitoring health status and performance.Performance— Maximize productivity regardless of I/O workloads, small or large files and I/Os, reads or writes, and random or sequential.The models:The HP StorageWorks X9300 Network Storage Gateway is a 2U rack-mount solutions that brings scale-out feature rich gateway file services to HP and third party arrays or SANs.HP StorageWorks X9320 Network Storage Systems are highly-available, scale-out storage solutions with performance and capacity oriented configurations.HP StorageWorks X9720 Network Storage System serves and stores hundreds of terabytes or petabytes of file-based data for expanding applications.HP X9300 Management Server is a 1U rack-mount solution used as the central management interface.
• #52: Notes to presenters: Key points:This is roughly the same slide as number 8. This is the last slide in the ‘Executive Pitch’. This is your opportunity to review the customer’s needs when it comes to unstructured data, and how HP solves those problems. This is also your opportunity to ask for any outstanding questions. And also, to set up the next meeting. The slides that follow this are not necessarily ones you would cover in an executive pitch, as they talk specifically about the X9000 features. However, the slides might become necessary. In any case, they are there for your use if needed.
• #53: HP CloudSystem, announced in January, provides a unified method for provisioning and managing services across private, public and hybrid cloud environments. The core infrastructure for HP CloudSystem is built on BladeSystem Matrix and 3PAR Utility StorageThe integration of HP 3PAR Utility Storage with HP CloudSystem centralizes the management of servers, storage and network resources under the HP Storage Provisioning Manager for the HP BladeSystem Matrix Operating Environment. As a result, clients can easily control a cloud infrastructure that spans servers, storage and networking via a single console. As businesses look to simplify operations and reduce IT silos this type of application-centric automation across the stack is a critical component and is why we are focused on unification of the entire infrastructure and not just a storage silo. Integration with Matrix reduces the time required for new cloud application deployment from days to minutesHistorically these CIO’s have had to decide between tightly vertically integrated solutions which sacrificed functionality across individual elements or loosely integrated vertical stacks cobbled together from best-in-class components.Now, customers can look to HP for leading innovation within each element of the stack as well as best in class vertical integration. 3PAR enables Matrix and CloudSystem customers to improve storage utilization and reduce costs 50% and simplify management by 90%.SOURCE INFORMATIONDeployment in minutes based on comparative calculation of setup and provisioning process for HP 3PAR storage integrated into the Matrix Operating Environment vs. manual cloud application deployments with heterogeneous infrastructure. Storage utilization and management efficiency improvements based on actual customer experiences with 3PAR thin provisioning and adaptive optimization technologies.
• #54: The StoreOnce D2D appliance automates the data protection process by providing high speed backup and recovery performance and enhances reliability with a highly available architecture. It also incorporates unique StoreOnce deduplication technology that was developed by HP Labs and delivers the most efficient data reduction in the industry. The benefits of StoreOnce can also be extended to disaster recovery and remote/branch office protection through the inclusion of bandwidth efficient deduplication-enabled replication. The combination of these advanced technologies and HP’s massive economies of scale creates a solution that is highly cost-effective and delivers a compelling ROI.
• #56: Slide Purpose: show the value of utility storage to hosted service providersOf the top 10 hosted services providers 8 use 3PAR as their hosting standard and one uses 3PAR as their internal IT standard. Service providers need a scalable flexible easy to management system in order to provide storage as a service to customer and be cost effective. The same values also apply to your internal IT.