VDI: The End of the Data Centralization Use-Case for Replication?

Admittedly, VDI is in the early days of adoption. As evidence, I cite TheInfoPro data shared at the recent Virtualization Summit at the Spring SNW event in Orlando: adoption of VDI will still be below 50% among the organizations surveyed by the end of 2010. As things VMware go, this is a slow growth area.

But, I have to wonder why. Organizations continue to struggle with remote data, both in terms of the cost of its management and the practicality of corralling it for protection. All kinds of fancy data replication schemes are in the market — some of which I have helped to promote over the years — for replacing backup systems in remote branches with replication to pull the data closer to the central management team. Replication products range from host-based to disk-embedded, and even fabric-embedded which is theoretically the most efficient placement for the replication engine.

In fact, centralizing data from remote sites with replication has become the second most prevalent use-case for replication. The first and most prevalent is disaster recovery, which I would actually argue is a combination of disaster recovery and making data available for offline backup. But centralizing data with replication from remote sites for backup purposes has gone mainstream in the last few years, having been integrated into backup applications themselves including Microsoft DPM.

With all of this momentum, then, why do I predict the end of this use of replication? Because VDI technologies solve the root problem better. VDI avoids the need to manage all of these in-bound streams of replicated data from remote sites, because with VDI the data never leaves the central site.

The beauty of the VDI model is that the technology has improved so much over the last couple of years, in particular with the VMware 6.5 version, that end-users don’t have to realize that their data is remote. With VDI, end-users have the same access and control of their files and information; but, the data is central and not remote which solves an entire range of management problems while lowering costs and risks. Think about these benefits:

• Data is easier to store, backup, archive, restore and retain
• Data is easier to produce for compliance and legal discovery
• Data is not exposed to theft of loss due to equipment theft and loss, such as when a laptop is left on a taxi
• Data is managed with the same more-robust infrastructure established for data centers
• Data is managed by better experienced IT teams
• Data can be stored on the same cost-efficient tiered storage established for data centers

Rather than having to replicate the data to the central point, with VDI the data serving remote end-users is already central and available for simple, low-cost management.

In the literature describing the benefits of VDI, I haven’t seen a lot of emphasis placed on the improvement that VDI offers in data management. Instead, I see much made of the benefits that VDI brings to simplifying the management and provisioning of desktop systems. A good benefit, certainly, but not one which tends to drive spending. Perhaps this explains why VDI is not catching the world on fire with deployments.Yet.

I think it’s just a matter of time, however, before IT teams catch on to the fact that VDI solves the remote data management problem better than do nothing, better than remote backup, and better than replication used to centralize the data for management. When that happens, replication vendors will need to be ready with new ways to add value to the data management problem.

Top 50: Cloud, SaaS and Virtualization - No Surprise There

For those of you who are interested in the ‘insider’ story, here are the “top 50″ startups to watch, as reported by Information Week:

http://www.informationweek.com/news/services/saas/showArticle.jhtml?articleID=216600068#

The Parable of the Wicked EMR…

I quite enjoyed this blog entry, on what can happen when medical records include - and propagate - medical errors:

http://e-patients.net/archives/2009/04/the-parable-of-the-wicked-emr-guest-post-by-david-kibbe.html
PS: Is it just me, or is EMR colliding with Social Media, and starting to pick up speed fast?

As VMware Momentum Continues, STORServer Reinvents Appliance Delivery Methods by Offering Encapsulated Virtual Machines

Among the vendor teams that I met at SNW this year, the team from STORServer stood out to me as having the best use of appliance technology built using VMware systems. STORServer offers an appliance built on IBM Tivoli Storage Manager (TSM) to simplify data backup. STORServer was there to talk about their VCB capabilities, which simplify VMware Consolidated Backup adoption - certainly a good use of an appliance solution approach. But more interestingly, I felt, was their use of VMware technology built into their appliance to speed the delivery of new applications capabilities to their customers.

Here’s how it works: STORServer offers new Virtual Machine configurations that can be added to the STORServer appliance already on-site in a customer environment. This can allow new turn-key applications supported by STORServer to be added to their appliance, without the need to purchase more hardware. This exciting applications-delivery method is demonstrated by STORServer’s new alliance with Mimosa Systems. Organizations can now get a preconfigured VMware-powered appliance that does both backup and archive: STORServer’s encapsulated TSM backup runs in one Virtual Machine, while Mimosa email archive runs in another.

The move by software providers to an appliance-based delivery model is something that I’ve been watching for some time, working with Bell Micro’s HP OEM team. I recognize the benefits of appliance delivery, both to purchasing organizations as well as to vendors: simplified installation, operation, support and maintenance make this delivery method particularly attractive in the SMB and commercial-tier market. Large enterprise organizations can benefit, as well, since every organization needs simpler-to-manage software.

For evidence of appliance success, consider STORServer’s backup appliance which encapsulates IBM Tivoli Storage Manager (TSM). Long recognized to be superior enterprise data management software, TSM has been too complex and expensive for most mid-tier organizations to consider. STORServer simplifies TSM by eliminating software installation, simplifying configuration through guided setup Wizards, and enhancing management with new control and reporting capabilities. Administrators using STORServer never see the embedded TSM software. But, they experience the benefits of TSM, including reliable, scalable backup capabilities like incremental forever, and efficient recovery.

Using this approach, STORServer has quietly captured over 500 customers and claims a 90% renewal rate which demonstrates customer loyalty. This business performance for an organization founded in 2000, while not a screaming success, shows that they have gotten something right.

Something else that STORServer seems to have gotten right: they work 100% through VARs, enabling VARs to participate in maintenance renewal revenue while they provide technical support direct to the customer. This approach provides VARs with the business incentive that they must have to help in the sales process, while enabling STORServer to maintain visibility to their customers.

Taking appliance capabilities one step further by building on VMware systems, as STORServer has done, seems an inevitable evolution which offers many benefits in flexibility and management. STORServer gets points for being one of the first. STORServer’s alliance opportunities seem wide-open with this approach, and their customers seem poised to benefit from a virtually endless stream of potential added capabilities (pun intended).

You can find out more on STORServer’s technology, at their web site: www.storserver.com.

Cloud here to stay?

Cloud computing has looked very ASP-ish to me. Which, to those of us who lived through the short-lived ASP craze that came and went in the late 90’s, is not auspicious for Cloud.

Here’s some evidence, tho admittedly shaky, that Cloud may have more staying power…

http://www.ft.com/cms/s/0/cc234f72-2e4f-11de-b7d3-00144feabdc0.html?nclick_check=1

Six Models of Cloud Computing Defined

I found this excellent definition of the six models of cloud computing, in my recent surfing: http://blogs.msdn.com/uspublicsector/default.aspx

Applies not only to federal space, but to all industries.

Why the Interest in EMR?

I’ve noticed the convergence of interest by end-users and vendors, on a few key topics which are: virtualization, cloud computing and healthcare. The ‘why’ of this is simple: there is money to be had in selling in these areas.

What’s perhaps unexpected, at least to me, is the strong convergence of how these areas overlap - meaning the selling of cloud computing, enabled by VMware, to healthcare organizations. You can see some of how this emerged at SNW from my notes from the Cloud and Virtualization summits (see www.dciginc.com for my blog entries on these topics).

What’s driving this convergence? It seems to me that the Electronic Medical Record, or EMR, is causing much of this. The promise to establish an EMR system was made by the Obama team during the primary season, and has been the focus of funding in the recent (Feb 2009) stimulus package. In Jan, in fact, Obama promised $50B in spending and targeted getting all records into electronic form within 5 years (see: http://www.edgeboston.com/index.php?ch=health_fitness&sc=health&sc2=news&sc3=&id=86262).

What is the need? Turns out that hospital systems cannot easily share patient records and information, with each other nor with physicians at point-of-care locations. This means that every time a patient is admitted to a new hospital or receives care from a new physician, he or she must repeat their medical histories, and even repeat expensive diagnostic tests and procedures like MRIs. Mistakes in medical histories account for a significant percentage of errors in patient treatment, some of which can be deadly such as when prescription medications interact.

What the Obama team recognized is that an online EMR system which keeps medical records up-to-date and available for all patients, no matter where they go for treatment, would save lives along with millions in unnecessary testing. This is a good idea. Getting it implemented will take potentially billions in new systems infrastructure, and radically change how hospitals need to manage data.

Turns out that putting the infrastructure for EMR systems into a cloud helps make patient record data universally available. Provisioning clouds with virtualization would make them more flexible and cost-effective.

So there we have it: the convergence of EMR, cloud computing and virtualization. Funded by billions of dollars in the 2009 stimulus bill.

SNW: Notes from the Cloud Summit

At this year’s SNW, I attended the Cloud Summit held on Monday afternoon, April 6. The speakers included Analysts and End-Users. Please enjoy my notes.

IDC predictions and data:

• Cloud adoption and growth are predicted to outpace growth of other spending over the next 3 years; CAGR of ~5%
• Cloud is expected to represent 25% of incremental new project spending by 2012
• $6.8B spent on SaaS in 2008; SaaS included in Cloud spending predictions
• 30% of organizations surveyed by IDC report using Cloud/SaaS in some form today
• 76% are working on active projects to adopt cloud computing
• The highest adoption rates for SaaS/Cloud are for CRM, HR, Collaboration Travel Expense and Sales Incentive applications

Key drivers for cloud adoption:

• Brisk pace of technology turnover and change
• High cost of capital
• Need to keep SLAs constant while staffing levels fluctuate
• Infrastructure spending this year is predicted to be relatively flat, up only .5% for all of 09
• Infrastructure spending is expected to increase in 2010 by 4.6%
• Worldwide infrastructure spending in 09 is expected to total of $367B for non-cloud/SaaS and $16B for SaaS; increasing to $438B and $42B respectively in 2010

Healthcare implications:

• Cloud is the ‘dream’ platform - perhaps the only platform - capable to deliver the Electronic Health Record/Electronic Medical Record (EHR/EMR) with universal hospital/healthcare access according to the CIO at Schumacher Group
• My own observation: EMR/EHR is the ‘killer app’ for Cloud Computing
• 70% of hospitals outsource emergency medicine
• Schumacher is the third largest and fastest-growing emergency medicine provider
• Schumacher uses Cloud service providers, all of which are HIPPA-compliant
• Schumacher’s observation: vendors of his on-site infrastructure are not as hospital-aware/healthcare-aware as his Cloud vendors

VMware implications:

• Virtual Machine technology (like VMware) is a necessary component for enabling Cloud application provisioning
• Enterprises are starting with Cloud secondary storage provisioning as a first step such as for data backup and disaster recovery
• Likewise, SaaS models for application services are well understood and relatively easy to embrace for enterprise teams
• Enterprises are less clear on how to use Cloud for production data analysis/provisioning