IBM gets serious about cloud computing

We've known for a while that cloud computing is important to IBM. It seems nearly every division has an effort in some aspect of the opportunity. And marketing has done its best to make it all look cohesive by wrapping these efforts under the Blue Cloud banner. But now we know they're serious. They have finally appointed a cloud czar, Erich Clementi, to bring all these efforts together. A veteran of their Systems & Technologies Group, he led SMB solutions, and last year he also took over IBM's Enterprise Initiatives. In those roles, Erich learned how important (and challenging) it is to coordinate efforts across their massive divisions while delivering holistic value to customers. Outside of Global Services, IBM doesn't have the best track record for these coordinated efforts, but we're willing to grant Clementi a grace period to prove us wrong.

And he won't be alone, as IBM has disclosed his set of lieutenants:

• Kristof Kloeckner from IBM Software, gets the CTO title and the task of defining and driving a cohesive architecture and strategy. He'll be charted with energizing the vast IBM R&D groundswell toward a singular direction. I think we can help you here, Kristof!
• Ric Telford is their product guy, trying to rationalize all the disparate SaaS related efforts across the company into an integrated (or at least coordinated) portfolio of services. This includes consulting and implementation services plus IBM's own SaaS offerings.
• Willy Chiu, also from IBM Software, will lead IBM Cloud Labs. Chiu runs IBM's HPC-focused HiPods lab efforts and has been banging the Blue Cloud drum the loudest of all with the ribbon cuttings on "cloud centers" throughout the globe (or are these just HPC labs with new shiny "cloud washing?").

The marketing and sales responsibilities will fall to:

• Mike Fay, from corporate communications and marketing, who will own the Blue Cloud banner and ensure message coordination and brand integrity.
• Mike Hill, from IBM Sales, who will drive customer engagements, the solutions built for them and Blue Cloud-branded sales programs.

No word yet on the goals for this new organization but knowing that one set of hands will be taking the wheel is an encouraging sign. I've described IBM's Blue Cloud efforts to date as a big listening engine. IBM spread the word early in 2008 through its massive marketing bullhorn that it was a player in cloud computing. They now are taking every opportunity to engage with clients to learn what they want from the cloud. IBM has studied how the clients themselves define cloud, taking that accrued knowledge with them as they build the proof of concept/solution that the customer has asked for. Now they have a central place to put all that information and a way to turn it into a coordinated effort.

We're in the early days of cloud computing, so Erich and team have some time to get this right. Most of their traditional rivals (HP, Cisco, and Dell) have had teams in place for a while now but it's not clear the competition (with the exception of Microsoft) are any farther along with a cohesive strategy. This is catnip for Salesforce.com and Amazon Web Services, who get even more time to extend their lead and build upon their substantial foundations. It's also music to the ears of the venture capitalists behind the many startups in the cloud computing space who gain time to build up their value as a potential accelerator for these large vendor's efforts.

Citrix and VMware have presented credible-sounding visions of cloud infrastructures that should see the light of day in the next 12 months, and Azure should lift the Technical Preview veil in this timeframe, so the clock is ticking for the major infrastructure players to get their act together from a competitive standpoint.

The real question, though, is, "Are enterprises ready to consume these cloud platforms and services?" Harvard Medical says it is. So does ESPN, Starbucks, and, of course, the New York Times. But are you?

We want to know. If your company is leveraging cloud services or cloud computing platforms today, tell us your story. We'll keep your confidence, if you request it, or help spread the word about your efforts via this blog and our exclusive client research.

Leave a comment below or contact us directly. Is cloud computing heating up or just the hype surrounding it?

By James Staten

Check out James' research

Knowing Isn't Hot But It Should Be

Last week Jason Newton at HP blogged about what his company thinks (or at least wants you to think) are the hot trends in the data center for 2009. He provides a good list that's less a reflection of what enterprise customers are necessarily doing but certainly what they should be thinking. Heck, his list reflects a lot of the tactics we discuss with customers every day in our inquiries and published research, such as in "Retrofitting Your Data Center for Better Capacity".

While you can certainly debate the priority of the items presented, there's very little here we'd disagree on (not sure DAS is HOT). At the center of all his advice though lies a subtle subtext captured best in his #3 - Knowing is HOT. Guessing is NOT. Newton writes, "Whether you’re talking capacity planning for your apps and virtual machines, the power and breaker size you need per rack or the storage for your data explosion, using the old ‘rules of thumb’ for quarterly budgets aren't going to cut it in 2009."

Absolutely right, Jason. But the issue is bigger than this. Enterprise IT Ops leaders are often flying blind on the overall cost of running the data center, what their average server utilization rates are, exactly how many servers they have, exactly how much storage they have and what percent of it can be used or shared. These baselines are necessary to truly know what level of positive effect you can have on your bottom line through consolidation, virtualization or any other efficiency improvement effort you plan to undertake. And this information is crucial if you need to deliver a business case to the CFO for investments in these efforts. And with all the consolidating and virtualizing going on today you'd logically conclude that, of course, knowing is hot. Everybody must be doing it.

Boy how we wish that were true. Unfortunately our data shows time and time again that most IT Ops leaders have a very limited understanding of their overall costs, let alone a complete asset inventory, utilization analysis or even an inkling of what portion of the power bill their consuming. Which means that most are making the case for their consolidation investments using the old "trust me" methodology. You know, "the old stuff is in efficient. The new stuff is far more efficient. So we can clearly save some money by moving to the new stuff."

Now I know that none of you are portraying your consolidation justifications this crudely, but without a baseline cost to work from, unfortunately a lot of the reasoning comes down to something like this. Vendors count on you taking this approach so you will use their ROI models, case studies and testimonials as your justifications. VMware likes to say that the average customers achieves 10:1 consolidations and has case studies showing millions of dollars saved. That's great. But will you really see those same results in your environment?

Don't take their word for it. Do the homework so you know exactly where your costs break down and which have the most acute inefficiencies. The best IT consultancies can conduct this analysis for you if you are slammed feeding the MOOSE. And a raft of good asset inventory and capacity analysis tools are out there that can not only help identify the baselines but serve as scorekeeper throughout the efficiency game.

Make Knowing HOT in your company so they savings you are chasing become a reality and IT becomes a real efficiency hero this recession, rather than a speculative one. 

Want help in making this happen? Come to our next IT Consolidation workshop, March 17th in San Diego, the day before our first Infrastructure & Operations Forum. 

By James Staten

Check out James' research

The Cloud, Not Deduplication Alone, Will Lead To The Demise Of Tape

Friday, Iron Mountain and Microsoft announced a new partnership. Customers of Microsoft's backup offering, Data Protection Manager (DPM) 2007 service pack 1, can electronically vault redundant copies of their data to Iron Mountain's CloudRecovery service. This is welcomed news for DPM customers. Customers will continue to backup locally to disk for instant restore but rather than vault data to tape and physically transport tape to an offsite storage service provider, customers will vault data over the Internet to Iron Mountain. For disaster recovery purposes and long-term retention services, you need this redundant copy of your data offsite. By eliminating the physical tape transport you eliminate the risk of lost or stolen tapes or the need to deploy some kind of tape encryption solution. Microsoft DPM hasn't taken the backup world by storm since its introduction in 2005, but each subsequent release has added critical features and application support. Additionally, because it is often bundled in with Microsoft System Center, I expect adoption will increase among small and medium businesses (SMBs) and small and medium enterprises (SMEs).

There are dozens of online backup services and new ones emerge everyday. I took the time note the Microsoft and Iron Mountain announcement because I think it's one example of a many partnerships to come between a commercial backup application vendor and a cloud storage service provider. Other examples include partnerships such as Zmanda and Amazon S3. Symantec has built its own "cloud" with Symantec Protection Network (SPN) and existing Symnatec Backup Exec customers can vault their data to SPN. Whether it's partnerships between commercial backup vendors and cloud storage service providers or vendors that build their own cloud storage target, these partnerships will help you eliminate physical tape.

Deduplication in backup software and disk targets such as VTLs has helped customers reduce their reliance on tape and for companies that have less data and shorter retention periods, to eliminiat tape. But deduplication didn't bring about the demise of tape as quickly as many thought. First, while dedupe certainly reduces the capital acquisition cost of disk (because you need far less disk), it's not removable so you still need some way of getting the data offsite, and that means you have to buy two disk systems, locate them at different sites, buy a license for replication, and potentially increase your bandwidth between sites. Not to mention, that until MAID (massive array of idle disks) technology becomes mainstream, spinning disk consumes a lot more power than tapes sitting on a shelf.

But now combine cloud storage with deduplication, and a far greater number of SMBs and SMEs can eliminate tape. Cloud storage services such as Iron Mountain's CloudRecovery service gives customers their alternate site and remote disk target without the capital investment in a second site and system and deduplication greatly reduces the capital acquisition cost of the local disk but also the bandwidth required to transmit the data. And of course, with the data offsite, not on your premise, it reduces your power and cooling costs and the footprint for backup infrastructure.

I'm interested in your feedback. Do you have any secruity concerns with online backup and cloud storage services? What are your adoptions plans for these types of services? Are you seriously considering them today or is your organization not quite comfortable with the idea?

By Stephanie Balaouras

Check out Stephanie's research

Riverbed Acquires Mazu Networks

Tuesday, January 20th, Riverbed Technology announced its acquisition of Mazu Networks for $25 million in cash with the possibility of an additional payment up to $22 million if future sales meet or exceed $35 million in the 12 months after closing. Mazu, which provides an application and device performance monitoring and reporting solution, will be integrated with Riverbed's current WAN Optimization offering to provide a more comprehensive solution for enterprises. In addition, the move also shows a competitive response to Blue Coat Systems' acquisition of Packeteer last year.

Mazu's history in network performance reporting and analysis makes it a great acquisition for Riverbed, further promoting its broader appeal to enterprises seeking this information for capacity planning, governance, and security needs. Additionally, as Mazu is a genuine innovator in application dependency mapping and application analysis, the acquisition instantly fortifies Riverbed as a credible player in application performance optimization. The full circle of application visibility and control is needed to fulfill this market need and now Riverbed has both phases of the cycle covered.

The analysis elements of application performance optimization remain as some of the most difficult goals in IT, but it is essential that we all find tools to improve this situation. Many tools can tell us we have a problem, but few can tell us why. While no vendor has yet conquered this enigma, Mazu offers technology that is among the more promising methods. Forrester believes this will serve Riverbed well and more importantly, its customers.

IT Service Management (ITSM) is all about the applications. Optimizing IT services requires us to optimize applications, not to simply report on them. The word “management” implies action. Reporting still requires action to be manual. Manual action must be minimized, so tighter linkages between data collection, reporting, analysis, and active controls are needed for optimization. As you build these linkages, you get closer to a vision of automating service adaptation. We are far from that utopian vision, but we inch closer with each link. The Riverbed-Mazu combination promises to connect several of these links.

The typical assimilation pains will delay the desired joint benefits of the marriage, as they do with all acquisitions. Still, this is a good fit for both companies. Riverbed needs improvements in the analysis phase and Mazu needs a) an active purpose for its analysis (i.e., what do you do with this information?) and b) a partner to give it better staying power in the market. As we stated, Mazu is a novel innovator, but too small. Current market conditions will prove hostile to smaller vendors. Customers may love the solution, but they need to know the vendor will be there for them in the future. This combination certainly helps.

In summary, the union of Riverbed and Mazu is a good step forward for application performance optimization. It is time for the industry to merge functions to deliver true optimization of applications. Control (Riverbed) is handicapped by a lack of guiding information. Analysis (Mazu) fails to accomplish its full promise without integration to a control platform. Together, they offer a much more compelling solution. Be patient, as this realization will take at least a year to materialize.

By Glenn O'Donnell

Check out Glenn's research

IT Compensation - No Slowdown Here

The Wall Street Journal reported yesterday that, while we continue to hear stories of layoffs, "consolidation", and dire financial straits IT pros are faring quite well. According to a study conducted by employment site Dice.com, IT pros saw an average salary increase of 4.6% last year over 2007. Whether that type of jump will occur throughout 2009 is up for debate and doubtful, it does seem that businesses are recognizing IT professionals for the work that they're doing to help bring efficiency and savings to the business. Good news for IT pros casting a wary eye on the road ahead in 2009.

I reported recently on the fact that spending in IT seems alive and well, based on research I conducted using data from our IT readership panel. Areas such as networking, virtualization, and consolidation are seeing budgets increase in 2009, all in an effort to continue the momentum of IT-drive bottom-line improvements.

So, the takeaway, IT is clearly one of the best places to weather this recession, as it's clear that business "gets" and is rewarding IT pros that can help save their businesses some cash, if not save them in full. So what should IT shops do that are concerned about staying on track?

• Don't shift investment focus. Key investment priorities cannot shift for I&O as a knee-jerk response to an economic downturn. Budgets and investment priorities should not shift in light of changing economic climates, in fact, doing so is likely to have negative effects.

• Look to efficiency technologies to extend gains. Your focus should be on technologies that bring efficiency to the business such as virtualization, consolidation of core networking, and creative solutions to extend the life of client-side assets.

• Communicate internal improvements that contribute to business efficiencies. As an IT leader, keeping front of mind the direct impact IT infrastructure and operations can have in helping to save the business and communicating this to members of the IT organization is a good way to allay fears of departmental staff reductions.

Keep and eye out for more detail on this theme in my upcoming research piece, "Making I&O Investments To Save The Business" publishing later this month. You can set up a research alert to be notified when it publishes.

[Ed. note] This post originally noted the Dice survey was conducted in partnership with the Wall Street Journal, which was incorrect. More information on the Dice Salary Survey can be found on the Dice site, here.

By Chris Silva

Check out Chris' research

Adaptive Infrastructure No Longer Just A Vision

At the beginning of this decade HP put forth a vision for the future data center that they have now fulfilled with both products and services offerings. Viewed by some at the time as a reaction to IBM Applications on Demand, HP coined Adaptive Infrastructure as its vision for a "composable" data center that let resources  be quickly and easily assigned to business services based on their needs and for IT Ops to achieve and maintain high utilization of their data center resources.

In 2008 HP started to fulfill this vision with the combination of its blade systems, Virtual Connect technology, Virtual Machine Management software, and the OpsWare-lite Insight Dynamics-VSE. It started customers down this path through HP Professional Services who debuted an Adaptive Infrastructure Maturity Model coupled with advisory and implementation services that plotted your course towards this reality. This month they added orchestration and recovery services to the Dynamics family and announced new AI Discovery centers, ProCurve software to enhance network orchestration, and enhanced services offerings to take advantage of these new portfolio elements. All these announcements add significant meat to Adaptive Infrastructure, which is one of the few visions you can actually implement today.

The timing is good for these announcements and for delivery against this vision as VMware, Cisco and others start to ratchet up their own marketing around a similar-sounding futureview.

Cisco debuted its Data Center 3.0 vision early last year with the introduction of its 10GbE Nexus products. It clearly sees the network convergence 10GbE will bring as a catalyst for its vision which speaks with a familiar ring about the orchestration and composability of resources. And with rumors spreading about a potential play on the server side, Cisco is garnering mindshare well ahead of its ability to deliver.

And last fall VMware launched its similar-sounding vision, Virtual Data Center OS. While VMware has yet to reach the market presence of HP or Cisco, it has the ear of many IT Ops leaders thanks to the IT consolidation and automation gains it has delivered to the market in the last four years. Its success has had such a strong effect that most vendors find themselves competing not with VMware's existing product line, but against their roadmap, as many of the IT best practices for virtualization implementation are being set by VMware.

While all these efforts at transforming the data center into a more responsive, highly utilized asset pool are commendable (and sorely needed), each carries a degree of proprietariness. Thus, IT Ops leaders need to ask: which vision should we follow and what tradeoffs do they carry with them?

In HP's case, you can't really deploy a heterogeneous Adaptive Infrastructure -- it runs on HP BladeSystem with Virtual Connect. You can use any hypervisor technology, any OS and can even incorporate your Integrity open systems into this model, but its capabilities quickly become limited off-platform.

Cisco's DC 3.0 obviously locks you into the Cisco networking products family. It will likely be best with VMware, based on their previous partnership announcements, and will likely attempt to commoditize the server and storage assets it orchestrates.

VMware's VDC-OS will clearly be a VMware-only (while it might incorporate other hypervisors) infrastructure play.

You'll be hard pressed to find any vendor's vision or strategy for IT modernization that is truly heterogeneous, so looking for that is a fool's quest. Instead look to your own standardization efforts. If you are standardizing on any of the above vendors, then you are likely to be inclined toward their vision and will be able to gain tremendously from what they bring forth.

If HP is your preferred systems vendor then you can have this vision now, which may be an important trump card if you are taking advantage of the recession by investing in transformation now so you can emerge stronger. Its one thing to have a compelling vision. It's another to be delivering it today.

By James Staten

Check out James' research

Nortel's Financial Firestorm: Implications For I&O Clients

Nortel’s future does not look bright. It admitted to huge financial reporting misstatements in 2002 and 2003, and the Canadian federal police are pursuing the former CEO, CFO, and comptroller for fraud. Nortel has lost more than $7 billion since CEO Mike Zafirovski took over in 2005. Competitors like Cisco and Juniper have been gaining share and new ones like Huawei have emerged. Since 2005, Cisco’s revenues have grown by almost 60%, Huawei’s have more than doubled, and we estimate that Juniper’s revenue has grown by well over 40% over the same three year period.  In contrast, Nortel’s revenues have declined.  During this time, Nortel’s share price has dropped by 90%. Nortel is losing share in the IPT market and the overall economic climate will clearly affect telecom hardware purchasing. All of this makes Nortel’s recovery even more difficult ¬— despite having $2.4 billion in cash on hand.

We anticipate that Nortel will continue to divest one or more lines of business.  For instance, it sold its GSM and 3G business unit to Alcatel in 2006. Some are speculating that its CDMA business unit could also be sold.  It will very carefully consolidate future R&D investments into products it sees as the most likely avenues for future growth. On the enterprise side, Nortel executives plan to continue focusing on Unified Communications, but on a call with Forrester analysts on January 15, they declined to provide specific plans or even timeframes for when those plans will be shared.

As a global “top five” provider in business and enterprise communications, Nortel has a sizable customer base that will not quickly disappear. However, if you are a Nortel customer, Cisco, Avaya, and Microsoft will probably be calling soon looking for your business.

If you are not currently using Nortel as a vendor, don’t start now. If you are considering a substantial additional investment in Nortel gear, hit the ‘pause’ button until its specific going-forward product strategy and product support strategy are clear.  If you are adding to your Nortel base in a small fashion — e.g. incremental phone or line card change outs, etc —  there’s no reason to stop. Continue to work with Nortel services partners for upgrades. Nortel’s extensive market presence and sound technology makes it highly likely that support and services for its installed base will continue through this difficult time.

If you are a customer on the enterprise services side and are considering using a new managed service that includes Nortel CPE, ask the MSP about other CPE vendors in its line-up. If you are planning to use a new telecommunications or WAN service whose underlying network is heavily dependent upon Nortel infrastructure, hit the ‘pause’ button until the service provider’s going-forward product strategy is clear. Finally, if you are using legacy services that use Nortel infrastructure, there’s no reason to jump ship.  But you should also question your service provider about its going-forward strategy about this part of its infrastructure. If you are considering switching providers for legacy services, make sure their RFPs include questions on underlying infrastructure vendors, financial strength, and going-forward strategy.

By Lisa Pierce and Elizabeth Herrell

Check out Lisa's research

Check out Elizabeth's research

Forrester Client Choice 2009

2008 was a tumultuous year for IT Infrastructure & Operations professionals. With an uncertain economic climate continuing to affect decisions in 2009, many companies have been forced to change the way they do business, scaling back on IT projects, staffing, and leveraging more outsourcing options. Clearly, we're not out of the financial woods yet, and IT is being counted on more than ever to bring change into the business.

Forrester knows the IT organization is no stranger to budget constraints and pressure from business to save money. It's been said that it's "always a recession in IT," and, to ensure our research is providing valuable insight on how IT is making decisions in 2009 Forrester's IT Infrastructure & Operations team wants to work with you to develop research topics that will answer your questions.

We'd like you to tell us what research could help you prioritize or launch 2009 money-saving initiatives? 

• Peer organization overview
• Server virtualization
• Green IT

Click on the "vote now" button below and see what your peers say will be their top money-saving indicatives for 2009, while helping us write the research that matters most to you!

Wi-Fi: The Only Network You Need?

CES, the showcase of all tech that is likely to make a splash over the next year, is upon us. While reserved primarily for, as the name implies, consumer technology, there are fundamental tools and technologies that will impact businesses as well.

News coming out of Las Vegas today is around chipmaker Intel's use of existing Wi-Fi radios for personal area network (PAN) connectivity. The release of the updated Wi-Fi driver software will allow the existing wireless 802.11 radios inside devices with the Centrino 2 chipset to use the radio as a means to connect to an IP-based network as well as local, Wi-Fi-enabled devices. The devices, ranging from cameras to network attached storage (NAS) will allow users to rely on the Wi-Fi radio in their Centrino-powered device to provide IP connectivity as well as file sharing and peripheral connectivity. Currently, the standard for most PAN applications and peripheral devices is Bluetooth, however Ozmo Devices, a firm partially funded by Intel, is out to change that.

While consumers may applaud the ability to connect to multiple resources while saving battery and not having to juggle dongles for various wireless accessories, the real appeal for making dual-use of the Wi-Fi radio in the device is its potential to advance the idea of the all-wireless workplace. Wireless infrastructure vendors have long talked about the future state of an all-wireless workplace where the LAN and resource connectivity is all wireless. Saving on the need for deploying wired network drops, the all-wireless story is appealing, if a suitable Wi-Fi network is already in place. Benefits can include a more flexible and more easily reconfigured workplace and the potential for lower power consumption due to having just wireless client devices to power in addition to printers and some access points - ideal for a branch office.

Intel has talked about the idea of its Cliffside technology - the ability to split the phsyical wireless radio in a client device into two separate, logical connectivity devices for the past year or more. Today, the ability to make this concept reality not only extends to new devices, but is backward-compatible with all 5100 and 5300 Intel WiFi Link chipsets.

What does this mean for an IT department contemplating an all-wireless deployment? Devices will not only to connect to the network using Wi-Fi, but can be connected to share peripherals and resources as they become available.  In addition to power consumption savings, the new multitasking features of the WLAN chipsets will allow for simpler standard configurations of mobile client devices, taking advantage of a newly multi-tasking 802.11 radio.

By Chris Silva

Check out Chris' research