Storage Peer Incite: Notes from Wikibon’s May 26, 2009 Research Meeting
Rising tides lift all boats, but falling tides smash some on hidden rocks and leave others stranded on sandbars, while the strongest and best positioned forge ahead, leaving the rest in their wake. Certainly this month's EMC World 2009 was a demonstration of which of those categories EMC belongs in. Just running a large and expensive event in a season in which other companies are canceling their conferences is a demonstration both of EMC's continued financial strength and its confidence in the future. And Joe Tucci's promise to continue fully funding the company's R&D shows both the depth of EMC's finances and its commitment to the future.
Despite early skepticism, some confusion and lack of substantive market traction, the 'Private Cloud' and Atmos are examples of how EMC is taking a long view at a time when others are just trying to survive. Cloud computing is the future, but it is not the present. A company that has the resources to invest heavily in developing new technology for what is today a very small market has both deep pockets and a vision. While there's lots of posturing with the Cloud, EMC is clearly positioning itself now for the future recovery, even if at present that recovery still looks a ways off. That is strength.G. Berton Latamore
Review of EMC World 2009
On May 18-21, members of the Wikibon community gathered in Orlando, FL, to participate in EMC World 2009. EMC World started as a grass roots effort by technology practitioners geared for technology practitioners. Over the years, EMC has layered senior management presentations, partner exhibits, and industry analyst and trade press material on top of the core technical content, to create the storage industry's most useful customer meetup. As EMC's business has evolved, EMC World has morphed into an information infrastructure event covering not only core storage, but also VMware, security, content management, and in the past two years, cloud computing.
The most striking aspect of the event is its size. In a climate where travel is down and other events are struggling, the organizers of EMC World and its attendees were not disappointed. EMC World 2009 had 7,000 attendees and more than 80 exhibitors and sponsors. While the number of attendees was off from last year's 9,000, it wasn't evident as the keynote sessions and technical breakouts were packed. The exhibitor hall was also very active as EMC's partners and suppliers were there to get in front the the storage industry's largest customer base.
On May 26, 2009, the Wikibon community met to discuss and summarize this year's event. What follows is a brief overview of the highlights of that call, which can be found at the Wikibon Peer Incite Archive.
Tucci and Maritz
The keynote presentations this year were kicked off by what are becoming EMC's #1 and #1A, Joe Tucci and Paul Maritz. It's clear that EMC's long-term strategy heavily involves VMware, and the company is intent on showcasing its value to customers, investors, and observers at large.
Tucci largely echoed his comments from EMC's March financial analysts meeting, in which he cited several growth areas including a somewhat expanded version of the big five, namely:
- Server virtualization - for every $1 of virtualization there are $3 of storage at the backend;
- FCoE which brings the Fibre Channel protocol to Ethernet, centralized management, and lower cost of connection;
- Cloud-based storage and private clouds;
- Data-center efficiency (aka Green) - through tiering and spindown;
- SSD and flash -- a total game-changer as price drops.
Maritz then took customers through the EMC/VMware vision of its private cloud strategy emphasizing that:
- EMC intends that VMware will become a platform for ALL applications.
- Performance generally and I/O performance particularly will not limit VMware adoption.
- VMware is building a software mainframe for the future.
- Its strategy is to encapsulate internal and external cloud computing infrastructures and allow customers a single management paradigm for security, compliance, and governance.
Notably, no customers we spoke with were planning to implement such a strategy in the near term, as much of this vision is in development and unproven. However most customers we spoke with saw the need for such capabilities, specifically as they pertain to so-called 'cloud bursting' -- the ability to accommodate surge demand for internal data-center resources by using the external cloud.
The Wikibon members in attendance at EMC World discussed the following observations:
Core storage - i.e. Symmetrix, CLARiiON, Celerra, Centera and data protection - the main theme is execution and EMC as the 'safe bet.' The company continues to spend heavily in R&D and expanding the functionality of its broad portfolio.
FAST - EMC's Fully Automated Storage Tiering solution, promised as part of the VMax announcement and implied to be available across EMC's core storage portfolio. FAST is a critical component for making storage more economically attractive. Few details have emerged as to how EMC will attack the opportunity to dynamically move data throughout the storage hierarchy and match data and device characteristics. Customers should push EMC for more details prior to committing major resources to VMax.
Cloud/Atmos - lots of talk but few customer references. The main emphasis is on developing the ecosystem to support object-based storage, but clearly there is a long way to go.
Content Management and Archiving (CMA) - According to Wikibon member Geoff Bock, Principle of Bock and Company - "CMA is launching a pincer movement to corral content in the wild. On one flank there's SourceOne with its emphasis on email archiving and records management; on the other there's Documentum with it's new xCP initiative for rapidly deploying content applications. Promising? Certainly. But the jury's still out and the devil's in the details." The key question is whether EMC can reconcile the fundamental flaws in its strategy between centralized document workflow and de-centralized risk mitigation.
VMware - EMC continues to demonstrate outstanding prowess in VMware integration stemming not just from its majority ownership of VMware but EMC's recognition of the importance of the platform and its ability to effectively mask the lack of array-based storage virtualization and continue to thrive in the market.
Iomega and consumer markets where EMC gets more disk drive buying power and a distribution channel into SMB. Iomega is now #2 and #3 in European and the United States, respectively. DAS markets behind WD and Seagate respectively.
Data De-duplication and data reduction. EMC's strategy is to extract intellectual property from its Avamar acquisition and place data reduction technologies (source-based, target-based, compression, single instancing, etc) across the portfolio and make them speak the same language. The benefits of this strategy are: 1) it will eliminate the need to 're-hydrate' data as it is moved through the infrastructure and 2) it allows data reduction to be both incremental and cumulative, increasing storage efficiency and optimizing data efficiency. It also is an effective attack strategy against companies that are primarily point-product based (e.g. Data Domain, although the NetApp acquisition changes that overnight).
Customers are advised to dig into this strategy and confirm that what EMC has delivered for file-based Celerra can be replicated in more complex block-based environments. Wikibon believes this is non-trivial and could result in de-duped silos.
Customer Loyalty, otherwise known as EMC's Total Customer Experience (TCE) program. Wikibon believes it is the most comprehensive and serious customer loyalty program in the storage industry and perhaps even the entire technology sector. EMC ties employee compensation to the results and is using TCE as a competitive weapon.
Notable Exhibits
In the limited time we had, Wikibon members focused on visiting a limited number of exhibitors including:
Axxana - demonstrating zero data loss at asynchronous distance integrated with EMC's RecoverPoint for CLARiiON.
QLogic, Emulex and Brocade - showing off FCoE solutions, with QLogic demonstrating a single-chip converged network-adapter (CNA) that was announced as shipping with several IBM servers.
Seven10 a provider of StorFirst Altus software, a Windows-based application which can replicate an EMC Centera archive in native-CAS-format to a VTL or a tape library for compliant disaster-recovery purposes.
Unisys - announced new ClearPath mainframes and integration with EMC's Symmetrix VMax as part of the offering.
Virtual Instruments - provides instrumentation software and services to probe large FC-based SANs in VMware environments. The key to the Virtual Instruments solution is allowing customers to identify bottlenecks and improve port utilization, thereby increasing VMware consolidation ratios, which are under pressure as VMware becomes more widespread.
ZettaPoint - showing a data classification engine that classifies data by usage and I/O to optimize database object placement, performance, and cost.
Note: Several times Wikibon members attempted to sit in on Cisco's UCS presentation, but it was consistently so packed we were unable to participate effectively.
Action item: In general, customer feedback at EMC World was extremely positive. While none were actively considering implementing a private cloud vision, customers expressed extreme loyalty to EMC as years of processes and skills built up around EMC's solutions are hardened. EMC World remains a highly useful event developed by practitioners for practitioners and for clients looking to maximize investments in EMC hardware and software it's a must attend annual event.
EMC: The strong survive in 2009
EMC has the strength to withstand the current economic downturn and represents a less risky bet for users. EMC's large portfolio brings with it lots of integration challenges, but with its continued emphasis on R&D investment, users can be assured of getting a rich set of features and functions. EMC's Joe Tucci made it clear in his keynote presentation to customers that EMC is not "letting up on the R&D pedal." In an effort to gain share in the downturn Tucci also indicated EMC wasn't pulling back on sales people-- again good news for customers, as EMC's culture empowers sales people to access resources to satisfy demanding customer needs.
Is EMC gaining share in the downturn? It's hard to tell. Small companies like Compellent have reported outrageous growth year-on-year but from a comparatively tiny base. 3PAR, another relatively small firm, announced flat revenues in Q1 (year-on-year) while NetApp's revenues were off 6% in total and 20% in product revenues (year-on-year) in its fiscal Q4, which most closely aligns to calendar Q1. EMC's "Information Infrastructure" revenue - where it counts the core storage business - was off in the low double-digit range year-on-year.
What's perhaps more important is that EMC is generating free cash flow (FCF) in the downturn. EMC's FCF was $681M in the first quarter and its FCF from the Information Infrastructure business was $494M. By comparison, NetApp generated $45M in FCF in its comparable quarter while the smaller firms are not even on the radar in FCF terms.
Given these outstanding FCF figures it's hard to argue EMC doesn't get richer in 2009. While numerous companies may be weakened by the downturn, the Wikibon community expects EMC's predictions that it will emerge from the downturn stronger will be accurate.
The key issue for customers is that if EMC is intent on gaining share in 2009, how can customers take advantage of this posture? One way is EMC's solutions business, which offers some of the best customer 'freebies' in the business. Users should take advantage of EMC's solution perks to reduce implementation risk by ensuring performance and configuration testing on EMC's nickel.
The other strategy users should consider is to approach 2009 EMC acquisitions as a project, organizing cross-functional teams with visibility into all EMC acquisitions across storage, security, content management, and services. This will ensure the highest quality service and best price from the storage giant.
Action item: EMC is strong and healthy, with bundles of resources, and loyal customers should take advantage of a buyer's market. Users should treat EMC negotiations as a project bringing together finance, legal, technical, business, and procurement teams with cross-organizational visibility. As part of this effort, customers should tap EMC's solutions group and secure free performance and integration testing for critical areas such as VMware.
Data Reduction Technologies (compression/de-duplication) key to EMCs data protection strategies
Percolating through the many presentations at EMC World was the notion that data reduction is a key strategy for addressing the ominous data explosion, which, according to Tucci, is expected to exceed 40% CAGR, even in today’s economic turmoil. EMC sees data reduction technology not simply as a part of storage efficiency but also as a significant component in its data protection strategy. This was nicely encapsulated in Tucci’s introductory comments that positioned data de-duplication as the technology that makes D2D back-up affordable.
The interesting twist in EMC’s perspective is the blending of its tiering strategy known as FAST (fully Automated storage Tiering) and data reduction technologies including compression, file level (single instance) and sub-file level de-duplication, both at the source and target side. The EMC vision is a move away from point products, such as Data Domain, and calls for the re-architecting of traditional back-up strategies and methodologies to enable the integration of data reduction technologies across the storage infrastructure. It is ambitious but if EMC can pull it off, will be impressive.
Avamar is EMC's lead product, and last year it had the distinction of being the fastest growing EMC product. Its IP has appearing in a number of products, with Networker announcing de-duplication on 5/19 and Celerra supporting a no cost, file-based de-duplication, available since February; both based on Avamar IP. My question is whether de-duplication is becoming commoditized. This would mean that its future as a point product is limited as it morphs into a standard array feature such as snapshot. This happens to be my perspective, and it explains NetApp's recent acquisition of Data Domain. Another possibility, perhaps supported by the recent EMC announcement of its data de-duplication assessment service, is that de-dupe will become a service.
During one of the breakout sessions the following questions were asked:
- Who is having bandwidth issues? No one responded.
- Who was meeting their BU windows? Mot one positive response was received.
When challenged, about the incongruity of their responses the audience had an “ah ha” moment. The bottom line is that a gap exists in the understanding of the full value of data reduction and particularly data de-duplication. The notion that source-based data reduction reduces the volume of data that has to move over the wire, and hence significantly reduce back-ups windows, is apparently not as obvious as often assumed. This attribute of source-based data de-duplication is one of its key advantages.
EMC has a strong vision, and despite proof-point successes such as Nationwide (which reduced its back-up window from 48 to 8 hours) and Celerra's file-based implementation, it will take some time before the results of this integrated approach will be apparent.
Action item:
- When exploiting data de-duplication as a standard BU solution, probably the solution of choice should be a point solution. Several options offer very competitive performance characteristics.
- If doing incremental de-duplication, source-based solutions would be the preference. This is where the EMC approach is probably strongest.
EMC World Cloud storage focus forces users to address organizing for the Cloud
Like it or not, the cloudburst of Cloud offerings such as EMC’s Atmos and Atmos Online make it mandatory for users to develop formal Cloud strategies. Not because of all the hype and media coverage, but because the technological innovations being developed today will be the tools of the future. Dozens, if not hundreds, of startups have sprung up with clever Cloud-based products. Many of those will be acquired by the larger players, and their technology will be promoted into the mainstream. Moreover, with EMC continuing to spend 12% of its revenues on R&D and its management focus on the Cloud, we can expect to see a bevy of EMC Cloud innovations, of which its recent announcement of the ability to federate public and private Cloud storage using Atmos is one example.
The first step in getting organized is to understand the type of project/business need before choosing a technology. Then identify the best fit based on the problem. Ask if the Cloud presents a line-of-business (LOB) solution or if it is an IT cost reduction strategy. If the later, responsibility for the technology likely belongs with the IT department. Many have found success in cloud computing by using the Cloud for surges in demand. One example is a company that offers e-mail scanning services via the Web. Most of its work is done in-house, but when demand exceeds capacity, work is shipped off to the Cloud, where it uses three vendors to process the excess work. Three were chosen to ensure application availability. In this example, internal IT has responsibility. On the other hand, if it’s more of a Web 2.0 competing service or a supply chain type of situation (e.g., Salesforce.com or Supplyscape) responsibility may be better off outside of IT’s hands.
Action item: Organizations should put in place a governance model to provide the parameters of the insourcing/outsourcing decision of Cloud applications based on the edicts of the corporation including security, availability, scalability, time-to-market, costs, etc. Do it now rather than later!
EMC and tiered storage; threat or validation?
EMC is developing an array-based, fully automated storage tiering solution called FAST, which will automate the movement and placement of data based on changing needs. Delivery is promised by the end of 2009 for all platforms. EMC has emphasized the role of flash storage in its storage hierarchy.
Other vendors are already shipping block array-based, tiered-storage solutions. The solutions from vendors such as 3PAR, Compellent, HDS, and Pillar Data are based on a fully virtualized architecture, where the data is split up into small pieces. This allows meta-data on parts of a volume to be captured and the data to be migrated. For example Compellent has already announced and shipped Data Progression™ on the Storage Center array, which automatically and dynamically classifies and moves data at the block level (within a volume) between storage tiers based on frequency of access. 3PAR and HDS offer management tools that monitor and move volumes between storage tiers dynamically. HDS offers the movement of data between heterogeneous arrays. In addition, IBM, EMC, and LSI are shipping appliance-based tiered storage solutions, again based on full virtualization of storage and management tools to migrate the data. All these appliance-based products offer movement of data among heterogeneous arrays.
A key business driver for tiered storage solutions is the growing disparity between cost and performance of different types of storage devices, from SLC flash storage at one end to high-density SATA disk technology at the other. The business case for tiered storage solutions is based on the better utilization of storage resources and the improved ability to meet service level agreements.
Wikibon suggests that an ideal block-based tiered-storage solution would have the following characteristics:
- Allow the definition of cost and performance characteristics of different storage tiers;
- Allow dynamic and non-disruptive movement of volumes and parts of a volume between different storage tiers;
- Allow movement between tiers across a heterogeneous group of arrays;
- Keep historical data on data access patterns and response times at the volume and sub-volume level;
- Provide a rules-based management interface for defining when data should be automatically migrated between tiers, taking into account time, access type, access patterns, and business importance of the applications supported by the storage;
- Allow a manual over-ride of automated data movement decisions;
- Provide management tools to allow the collection of data for monitoring data movement and data performance for use in reporting, charge-back, and storage capacity planning.
Action item: With validation from EMC, automated tiered storage solutions including flash will turn from a nice-to-have to a basic requirement in 2010. Vendors will need to have their tiered storage solutions ready and tested by then, or they will be drowned by the marketing noise from EMC. Users should start defining their requirements and understanding the business case for their environment, in preparation for assessing and implementing solutions in 2009 and 2010.
EMC World and EMC FAST
At EMC World, Joe Tucci said that solid state disk (SSD) is a game changer for the industry. Further, EMC has announced its Fully Automated Storage Tiering (FAST) technology, promising that FAST will be available across all its primary storage platforms by the end of 2009. We now live in a storage world where having multiple types of online storage devices, from SSD to SATA disks, makes sense. This is because it allows us to put the critical, current transaction data on the highest performing storage when it is needed and keep less-frequently accessed data on cheaper, somewhat slower storage. As FAST becomes a reality, it will enable storage administrators to automatically move data to an appropriate location, either within a disk array or to different disk arrays, at least among the EMC product family. This becomes especially important as automated server virtualization management and movement functions begin to appear.
As SSD pricing continues to drop, we can expect to see a change over time in the way we configure online storage systems. We may find that the most mission-critical data can be kept on a small amount of SSD storage and the large majority of data can be kept on slower, SATA storage. This two-tier online storage management strategy is relatively simple to understand and fits into existing management strategies, leaving room for other tiers for archiving, backup, etc.
Having the ability to automatically move data across multiple types of storage arrays is compelling, and EMC has stated this as its direction. The next step will be for customers to identify or classify data appropriately to take advantage of this automation. FAST works with all types of online storage, including SSD and all types of spinning disks.
Action item: EMC customers should look for opportunities to reduce storage costs by identifying data that would make sense to put under FAST management when it becomes available. VMware customers should plan on taking advantage of FAST capabilities as they implement automated VMotion functions.