Cloud Is Defined, Now Stop the Cloudwashing

This blog post is a response to an article by Alex Williams on ReadWriteWeb. Thanks for the shout out, Alex, and for bringing more attention to the contentious issue of cloud computing definitions. While Forrester research reports are created exclusively for our clients, our definition is freely available:

A standardized IT capability (services, software, or infrastructure) delivered via Internet technologies in a pay-per-use, self-service way.

We first published this definition back in March 2008 in the report, “Is Cloud Computing Ready for the Enterprise,” and have held to that published definition ever since (in fact it has been leveraged in multiple Forrester reports, speeches at industry events, news articles, blog posts and tweets since that original publication). Our definition was also used by NIST and several other Federal government agencies as a resource used to create their definition.

One of the key values Forrester Research provides to its clients is helping them navigate the technology trends and delineate what is a new type of technology and what is simply last year’s technology in new clothing (what I call “cloud-washing”). Thus it was imperative for us to publish our definition early and we have since been striving to provide clear taxonomy and categorization of cloud services, as shown in our latest Forrester Tech Radar.

In this Tech Radar we lauded NIST for their definition, and contrary to your statement, do not believe we need the “circus” of more or better definitions at this stage. Rather we believe we need broader recognition of what is and what isn’t cloud computing to get past the marketing hype and make it easier for customers to identify and then consume these valuable new service offerings. That’s why we’ve stuck with our definition since 2008 and are glad to see NIST sticking to theirs.

By James Staten

Check out James' research

You should follow James on  here

Cloud Computing Belongs On Your 3-Year Roadmap

Welcome to the fourth quarter of 2009; what we at Forrester call planning season for most IT departments. In a typical year, this is the time that infrastructure and operations professionals spend lots of cycles burning through what remains of the 2009 budget and building plans for investment in 2010 with the hope of gathering a bit more budget than last year. Of course this is no ordinary year. Economists and financial prognosticators, like our own Andrew Bartels are predicting a long recovery from the recession and further delays in IT spending. That means another year of your infrastructure getting older. There’s two ways of looking at this problem and thus your budget proposals for 2010:

• You can either predict when you will get the rights to refresh the systems and return to the infrastructure spending patterns of old, or

• You can realize that a long recovery means a new normal for IT spend is in order.

We think the latter is far more fruitful even if it isn’t entirely accurate. It’s both prudent and different because your go-forward infrastructure strategy likely isn’t the same anymore either. If you are like the respondents to our Q3 2009 Enterprise and SMB Hardware Survey, North America and Europe, you:

• Continue to consolidate infrastructure like a mad clear-cutting logging company.

• Are repackaging workloads into VMs faster than an Amazon.com shipping manager.

• Are standardizing like McDonalds – everything from hardware, golden master server images to change management processes as consistent and repeatable as possible.

• Are racing like Usain Bolt to shorten deployment times from weeks or months, to new world records.

• Are learning to love automation – if only in a small way.

In other words, you aren’t buying, running or managing infrastructure the way you used to. So why plan your budget the old way. And further, why budget for the new now, when you can budget for what’s next in your journey. As we noted in our report, Assessing Your Infrastructure Virtualization Maturity, there is a clear path of improvement for infrastructure architecture and that’s toward customer empowerment, deployment standardization, provisioning and capacity management automation and further consolidation – even across business units. Yes, I said it.

In other words you’re evolving into an Infrastructure as a Service cloud. But despite vendor claims for products such as the VMware vCloud, Eucalyptus and Elastra offerings, most enterprises can’t just drop in an IaaS platform and they have a cloud. You have to prepare your organization and your operations staff for this. But that doesn’t mean you can’t achieve this aim in a short time horizon or that you can’t embrace public or hosted cloud infrastructures in the near term.

Forrester feels that cloud computing is one of the Top 15 Technology Trends and that it warrants investment now so you can gain the experience necessary to take advantage of it in its many forms to transform your organization into a more efficient and responsive service provider to the business. Find small non-critical projects to start with so you can learn how best to apply these services to your business and combine this learning with the advice in our Tech Radar to help plot the timing for these investments. Also read through the many cloud computing case studies like those on USA.gov and CryoPort to see how others are doing this. By investigating these top 15 technology trends and shifting to a strategic rightsourcing approach to IT portfolio investment you can hollow out the MOOSE and become more strategic (and viewed more as a change agent) to the business.

By James Staten

Check out James' research

You should follow James on  here

The Role Of IT Operations In Archiving

Yesterday IBM announced the availability of their new IBM Information Archive Appliance. The appliance replaces IBM’s DR550. The new appliance has significantly increased scale and performance because it’s built on IBM’s Global Parallel File System (GPFS), more interfaces (NAS and an API to Tivoli Storage Manager) and accepts information from multiple sources – IBM content management and archiving software and eventually 3rd party software. Tivoli Storage Manager (TSM) is embedded in the appliance to provide automated tiered disk and tape storage as well as block-level deduplication. TSM’s block-level deduplication will reduce storage capacity requirements and its disk and tape management capabilities will let IT continue to leverage tape for long-term data retention. All these appliance subcomponents are transparent to the IT end user who manages the appliance – he or she just sees one console where they define collections and retention policies for those collections.

It’s a solid (albeit a few years late) announcement from IBM. As a leader in content management and archiving software, it was always puzzling that IBM lacked an attractive archive storage platform to complement these offerings. In addition, archiving is still a priority for both the business and IT. Business leaders remain concerned with archiving data appropriately for regulatory compliance and legal discovery. IT struggles to keep up with storage growth which continues to grow at 30%-40% each year and to control storage costs. IT has always wanted to archive infrequently accessed data from expensive production storage platforms to less expensive archive disk and tape but their hands are tied unless business owners, legal and knowledge management professionals define appropriate retention policies as well as their functional requirements in a solution.

Which leads to a question that I’ve always struggled with myself? What is the role of the IT operations professional in archiving? It’s clear that IT operations professionals are not responsible for defining the actual retention policies but what are IT’s responsibilities and how can IT influence business owners, legal, internal auditors and knowledge management? I’m still fleshing out the list responsibilities but here’s what I have so far:

1. IT can recommend a retention policy for certain data sets based on an analysis of the storage environment. For example, if you’ve examined your environment and you’ve determined that users have not accessed 60% of your files in two years, these file are good candidates for archiving.
   
2. IT researches, shortlists, and participates in the evaluation of archiving software or service based on the functional requirements that business, legal, audit, and knowledge management professionals define. This could include the evaluation of on-premise solutions, cloud-based solutions or a hybrid of both.
   
3. IT researches, shortlists, and selects the appropriate storage archive to physically store the data. This could be a generic disk storage system or a purpose-built storage system and it also includes tape and even cloud storage services.
   
4. IT defines the architectural requirements, scalability requirements, and performance requirements of the archiving software or service. At the end of the day, IT operations will manage the archiving solution end-to-end. IT operations will be responsible for maintenance of the archiving software, the database, operating system and the virtual or physical server that it runs on and of course, IT is responsible for the disk and tape that physically stores the archived data. This is one area where IT operations must really exert itself influence and its own requirements. Too often, performance, scale and system manageability are not taken into the selection of the archiving software or the determination of the total cost of ownership. When a software solution doesn’t scale or perform IT ops answers the calls from the help desk and is on the hook to solve it.
   
5. IT is partly responsible for maintaining the long-term accessibility of the data in the storage archive. This issue here is the backwards compatibility of tape formats. Tape formats such as LTO are only backwards read compatible with the previous two tape format generations. This means that you’ll need to periodically recall tapes from the deep archive and migrate the data to the latest generation. This address tape media deterioration and tape format obsolescence but it doesn’t address the obsolescence of the application that generated the data. Call the knowledge management team for that one.

This is just a starting list. I’m interested to hear from other IT ops professionals if I’ve missed something or if you disagree with anything on my current list. I’m also interested to hear the struggles you have with archiving.

By Stephanie Balaouras

Check out Stephanie's research

You should follow me on  here

When is a VMware cloud a vCloud? Let the API wars begin

Attention enterprises —Pop Quiz: If your favorite hosting provider launches a cloud service that supports VMware vSphere and is part of the VMware vCloud initative, are they providing you with the rich vCloud functionality VMware is touting at VMworld this week?

The answer is no, because vCloud hasn’t been delivered yet and probably won’t be until the first half of 2010. If you answered yes, don’t feel too bad because VMware, in its efforts to prepare the market for vCloud, has made the current state very confusing. First off, the company grandfathered all the members of its VMware Service Provider Partner (VSPP) program into being vCloud partners. This has led many a traditional hoster – who doesn’t even have an infrastructure as a service (IaaS) cloud platform to say they are now a cloud. Sorry, but if you can’t automatically provision VMs upon request from a self-service portal, don’t have pay per use billing and require a 12-month commitment, you’re not a cloud platform. Second, a hosting provider can support the vSphere VM format without letting you manage your VMs in a standard way (they don’t even have to give you access via vCenter). And third, if you want to integrate with their platform programmatically, their APIs are likely their own.

Why does this matter? You use VMware in-house, you just want a cloud that does too. The VM format supported isn’t really the issue (heck converting your VMs from VMDK to OVF, Xen, or Amazon AMI is fairly trivial. And tools like Citrix’ Project Kensho do this very effectively).  Management is where this really matters because the simplest way to consume IaaS cloud resources is when you can use your own tools for deployment, monitoring and reporting, administration, security, integration and life cycle management. And that simply isn’t the case today (thank you Amazon for letting enterprises finally manage cloud security with their own tools).

The same criticism can be levied at all Xen-based clouds today, too. Amazon, Rackspace, GoGrid and several other hosters have published the APIs to their cloud platforms but only Amazon’s, thus far, have a realistic chance of reaching de facto standard status (especially when Eucalyptus and the Sun cloud – if it ever sees the light of day under Oracle – use the same APIs).

Interoperability requires consistent APIs – a subset at least – among cloud platforms. And VMware took this step today. It’s a pre-emptive move and a smart one to release its vCloud APIs and hand them over to the DMTF to manage as public APIs, because it lays the groundwork for a promise of better interoperability and compatibility between IaaS clouds (and enterprise internal and hosted clouds) when vCloud finally arrives. To try to ensure that their APIs can deliver this value, VMware also announced its vCloud Express program that asks for a commitment from cloud hosters to support and expose the vCloud API within their API set ensuring not only a clear conveyance through marketing that this cloud is compatible but giving the hoster the ability to differentiate around this API by intermingling it with their own unique functions.  This is a good move by VMware because this guidepost will be key for enterprises that want a cloud provider that can most closely match their in-house deployment. It also gives ISVs some runway to build their solutions and have them ready by the launch of vCloud in 2010.

It’s also competitively important in light of the momentum Amazon Web Services has with its APIs and this week’s announcement by Xen.org to raise its level of cloud compatibility.

This announcement has also moved the hypervisor wars to a new battlefield – programmability. Because at the end of the day, it’s less about how good your virtualization solution is (there’s a clear bar of function, performance and maturity that must be crossed – and XenServer and VMware both are over that bar); it’s how big your market share and how robust your ISV and hoster ecosystems are that matter. Xen has clear hoster ecosystem market share leadership in IaaS clouds. VMware brings the muscle of its enterprise and ISV market share to this battle.

The war is really just a skirmish today as still just a fraction of enterprise buyers are leveraging IaaS cloud services. Forrester survey data shows that while less than 10% of enterprises are leveraging IaaS clouds today, more than half plan to in the next 2-3 years so there’s a lot to be gained from victory. 

  

Let the API wars begin!

By James Staten

Check out James' research

You should follow James on  here

Xen.org: Stop the Forking. Forrester: About Time

 Xen.org, the open source community behind the leading IaaS cloud computing hypervisor finally made a bold move today by stepping up to the plate of delivering a complete open source virtual infrastructure for cloud platforms. Prior to this release, Xen.org had been content to manage and maintain the core Xen hypervisor and let its partners all build solutions around it. The problem with this approach was that while the hypervisor itself was compatible between these solutions the infrastructure and how you managed it were not. Compatibility at these higher layers is what is necessary for the Xen community to attract an ecosystem of ISVs and end customers – no one wants to have to tune their software for each and every Xen variant. Now Xen.org is delivering a complete product that IaaS clouds can more easily deploy, Linux distributions can deliver and that end customers can install knowing that storage, deployment, monitoring, reporting, policy-based automation and most other management tasks are consistently executed. This also makes Xen plus Eucalyptus a more complete cloud-in-a-box, open source solution.

Xen.org even went so far as to make the official Linux kernel the domain 0 for this solution, addressing a contentious point that rival hypervisor KVM had raised as its differentiator. Xen.org vendor partners can still choose to leverage just the hypervisor for integration with their own infrastructure and OS kernel technologies (as Sun has done with xVM) but there will be diminishing value in doing so with products aimed at the mainstream market, as compatibility will be broken at these higher levels. Oracle is one worth watching as its OracleVM uses Xen plus Oracle Unbreakable Linux, which is based on Red Hat Enterprise Linux server. Red Hat is unlikely to add the full Xen infrastructure as it has shifted to KVM as its strategic hypervisor. Could this be the opening Oracle has been waiting for to break away from Red Hat with its Linux (Is Unbreakable SuSE on the horizon)?

Clearly this move benefits Citrix’ aims as well. The company carrying the Xen.org flag will be contributing all its cloud-relevant technologies to this open source project including its StorageLink, virtual switch, virtual appliance, and VMware-to-Xen VM conversion (Project Kensho) technologies. Many early IaaS cloud leaders have their own solutions in these areas but are likely to welcome this IP so they don’t have to maintain their unique solutions long term.

Hopefully with this move the days of cloud incompatibility, at least between Xen-based clouds, will come to an end just a bit sooner.

By James Staten

Check out James' research

You should follow James on  here

Cloud DR Services Are Real

There is a lot of hype surrounding cloud and I'm usually not one to join the hype but in the case of cloud-based backup and disaster recovery services (I'm trying to use the IT service continuity but it hasn't caught on yet), these service are available today and they address major pain points in IT operations and organizations of all sizes can leverage these services, not just small and medium businesses.

Storage-as-a-Service is relatively new. Today the main value proposition is as a cloud target for on-premise deployments of backup and archiving software. If you have a need to retain data for extended periods of time (1 year plus in most cases) tape is still the more cost effective option given it's low capital acquisition cost and removability. If you have long term data retention needs and you want to eliminate tape, that's where a cloud storage target comes in. Electronically vault that data to a storage-as-service provider who can store that data at cents per GB. You just can't beat the economies of scale these providers are able to achieve.

If you're a small business and you don't have the staff to implement and manage a backup solution or if you're an enterprise and you're looking for a PC backup or a remote office backup solution, I think it's worthwhile to compare the three year total cost of ownership of an on-premise solution versus backup-as-a-service.

If you lack a recovery data center or if you're looking for more a cost-effective solution for your Windows and Linux x86 server farms, I recommend that you evaluate new virtual recovery services. And for anyone who is about to renew their DR services contract, you should evaluate these new services before you resign for another three years of tape-based recovery to oversubscribed, shared IT hardware. These services fill the gap between DR services that rely on replication to dedicated IT hardware which are far too expensive for most organization and tape-based recovery to shared IT hardware, which while affordable, do not meet the recovery time and recovery point requirements of organizations from small to large and acoross industries.

And of course, if you're focused on the availability and recovery of email in particular, well there is a laundry list of service providers that focus cloud-based services for email from including continuity of service, recovery, archiving and if you wanted to, just move your email to the cloud period.

These services already have a toehold in the US but in order to expand into Canada and to Europe, these service providers are busily opening up data centers in these regions. In Canada and the EU, data privacy laws restrict the transfer of data containing personally identifiable information out of region or to a country that doesn't meet the standard of its own privacy laws.

Perhaps I'm too bullish on these services, perhaps organizations still have security concerns with cloud-based services but I think evey IT operations professional must now seriously consider these services when evaluating any backup or DR offering.

By Stephanie Balaouras

Check out Stephanie's research

You should follow me on  here

HP Acquires Cluster File Storage Software Vendor IBRIX

Today, HP announced that they will be acquiring the startup cluster file storage software firm IBRIX in a deal with few disclosed details. Congrats to HP for making another interesting acquisition in the clustered storage space, following their Feb. 2007 acquisition of Polyserve and their October 2008 acquisition of Lefthand. IBRIX makes software that allows users to cluster many nodes of storage (either industry standard servers or iSCSI/FC SAN arrays) for use in high performance global namespace file storage.  IBRIX’s customer base includes some 175 firms, generally in the high performance compute (HPC), video rendering/animation, and oil and gas exploration spaces.

This is an interesting move for HP who points to building share in HPC, content depot/cloud storage and archiving as key goals for this acquisition. Clearly those are high growth markets, so it’s no surprise that HP is interested in accelerating their traction there. But, it is surprising to see archiving and content depot listed as top level goals given that IBRIX doesn’t have WORM capability or object orientation. Additionally, you have to wonder what this says about Polyserve given that they were a maker of high scale global namespace file storage, and is now the basis for HP’s current cloud offering, XDS. There seems to be some overlap there, and it will be interesting to see what lines they draw to delineate use cases for customers.

Another area of concern here is the significant partnerships that IBRIX has with HP rivals EMC and Dell.  IBRIX was the original software used for EMC’s HULK scale out hardware offering before they completed the MAUI software that combined with HULK to make Atmos, EMC’s cloud storage offering. There are still a large number of users with highly complex environments that have IBRIX software mated to EMC as well as Dell hardware as back end storage; HP will have to be careful to avoid disrupting those power user customers as they develop their strategy for moving forward with IBRIX.

As with Polyserve and Lefthand, IBRIX is software only, so it makes sense for it to be owned by a major hardware player that can sell it as a comprehensive solution including hardware and services, rather than having customers cobble solutions together themselves. But, it will be tricky for HP to determine how to support existing solutions with gear from competitive vendors, how much to combine it with other software versus leaving it as a standalone product, and how it can fit with HP’s broader vision of storage.

On the positive side, HP has many back ends that could fit different types of IBRIX deployments, from Proliant servers to Lefthand iSCSI storage to EVA mid range SAN arrays, so the likelihood is that they have the right raw materials to make it work. On the negative side though, HP now has a very broad lineup of storage products, some internally developed, some acquired and some delivered through partnerships, and maintaining a consistent vision and value proposition that meet specific customer requirements may be a challenge.

By Andrew Reichman

Check out Andrew's research

Deduplication Market Undergoes Rapid Changes

In May, I blogged about NetApp's announced acquisition of deduplication pionneer, Data Domain. The announcement triggered an unsolicted counter-offer from EMC, followed by another counter from NetApp. But after a month of offers, counter-offers and regulatory reviews, EMC ultimately outbid NetApp with an all cash offer of $2.1 billion. I believe that Data Domain would have been a better fit in the current NetApp portfolio; it would have been easier for NetApp to reposition its current VTL as a better fit for large enterprises that still planned to leverage tape. It's also said that more than half of Data Domain's current employees are former NetApp employees so there would have been a clear cultural fit as well.

For $2.1 billion, EMC gets Data Domain's more than 3000 customers and 8000 installs but it also gets a product that in my opinion, overlaps with its current Quantum-based disk libraries, the DL1500 and DL3000. In Forrester inquiries and current consulting engagements, Data Domain is regularly up against the EMC DL1500 and DL3000. EMC will need to quickly explain to customers how it plans to position its new Data Domain offerings with its current DL family, both the Quantum- and Falconstor-based DLs as well as its broader data protection portoflio that includes Networker and Avamar - which also offer deduplication.

In addition to the EMC and NetApp stuggle to own Data Domain, we also saw one backup application vendor after another add deduplication capabilities to their software. The online backup vendors came forward to remind everyone that they have been deduplicating data in software all along. As I completed my current state of the union report on the deduplication market, no sooner had I completed one research interview when another backup vendor would announce new dedupe capabilities. I could have delayed the report forever waiting for a lull in the announcements. I officially interviewed 19 hardware and software vendors for the report but was briefed by an additional 3 vendors in the past 2 weeks. I already know there will be additional vendor announcements in the coming months.

It was also clear as I was writing the report that the definition of many deduplication characteristics are blurring such as inline vs. post-processing and source vs. target. I took hard stances on some of these definitions and forced vendor offerings into one category or another for the sake of clarity but I also tried to explain the nuances of some of the vendor offerings. For IT professionals, at the end of the day, you must understand essential deduplication characteristics and considerations and familiarize yourself with all the vendor offerings in the market but at the end of the day, you must have well defined performance, capacity, scalability, manageability and resiliency requirements. You must issue an RFI and/or RFP, let the vendors make the case for their offerings, short list 2-3 vendors and bring them in for a proof of concept (POC) to validate their claims. Refence customers and POCs are the best way to cut through all the hype, claims, and counter-claims made by the vendors in this market.

By Stephanie Balaouras

Check out Stephanie's research

You should follow me on  here

Planning to Upgrade Your IVR? Consider the Choices Between Premise Based and Network Based Solutions.

According to the recent Forrester Enterprise And SMB Networks And Telecommunications Survey, North America And Europe, Q1 2009, 32% of the 279 network and telecommunications managers surveyed indicated they planned to upgrade their IVR in the next 12 months. Before a decision is made, companies need to consider their options for upgrading their IVR and compare the differences between premise based and network based voice portals. Voice portals are standard based platforms that support multiple speech or touch tone applications. Forrester’s survey indicates 22% of companies plan to add speech applications this year to improve automation of customer transactions and provide better customer service.

Look For the Solution That Best Fits Your Business

Network based services provide a means to support advanced applications from a carrier’s network. The benefits of network based services are flexible OPEX pricing and faster time to deploy new services. Forrester’s Wave on network based IVR/voice portals provides an in-depth look at major vendors offering these services and insight into their value added solutions. Premise based solutions also support advanced VXML speech applications on native-SIP platforms. These standards-based platforms support innovative multimodal applications and provide Web service for integration with third-party cloud services. Organizations that want greater control over their application find premise based solutions better suited for their business. Forrester’s Wave on premise based IVR/voice portals has a new line up leading vendors in this area.

By Elizabeth Herrell

Check out Elizabeth's research

Green Storage Has Limited ROI, But Supports Overall Efficiency

I care deeply about the environment, certainly more than I care personally about money, so it pains me to say that in most cases, making storage decisions based on power expenditure alone is not rational behavior. The world is driven by economics, and the stark reality is that the cost of power is only a drop in the bucket compared to the amount organizations spend on acquiring and managing their enterprise storage systems. Maybe someday a consumption tax or cap and trade system will tip the balance towards more responsible consumption of non-renewable resources, but in the meantime, the pricing of power (especially in the US) doesn’t give much economic incentive for good behavior. In fact, according to a report Forrester published recently, the amount of money typically spent on electricity to power and cool a TB of storage is only about 1% of the cost of buying that TB of storage (or about 4% of the annualized cost of buying that storage given that you only have to buy the TB once every 3-5 years but you power it every year). So, unfortunately for the environment, power cost itself doesn’t provide a very strong incentive for storage efficiency.

Fortunately though, the things that enterprises can do to reduce their power consumption costs are often the exact same things they can do to reduce the capital and operating expenses of their overall storage environment. Focusing on improving utilization (measured as the quantity of data written divided by the quantity of storage on hand) and increased usage of dense drives are the most straightforward and effective ways to reduce hardware acquisition costs as well as power consumption. There are many ways to achieve these objectives such as thin provisioning, reporting and reclamation to improve utilization, and tiering or wide striping to enable more use of dense drives. Whatever the motivation — economic, altruistic, or a combination of both — organizations that put significant focus on their utilization and dense drive ratios are likely to spend less money and be greener at the same time. And that’s good for everybody.

By Andrew Reichman

Check out Andrew's research