CA Community

October 2010 - Posts

GITEX 2010: Clouds in a country where it never rains

Published: October 27 2010, 03:26 PM | 1 Comment(s)
by CA Community

Last week Jim Murphy of Gartner and I opened up the Cloud ConfEx, an executive conference that ran as a part of the GITEX tradeshow in Dubai.  I had not been in the region for the good part of 10 years, so I was very interested to see how it has developed.  Last time I was in Dubai, the Burj Al Arab was there and so were some of the other landmark buildings.  But the artificial islands called The Palm and the tallest building in the world, the Burj Khalifa, did not even exist as an idea back then.  I went to Dubai with questions about how high the interest in cloud computing would be and what its potential is in an emerging region and growing economy like the United Arab Emirates.

The initial feedback from the market research we are completing on cloud computing in emerging markets (to be published in a few weeks) shows some reluctance and apprehension towards private cloud, but a significantly higher interest in public clouds.  Jim Murphy gave Dubai a 24-hour head start on the rest of the world by presenting the Cloud approach that Gartner launched to the rest of the world at its ITxpo event in Florida a day later.  Prior to Jim's presentation, I shared with the audience my thoughts on how cloud can be especially beneficial for doing new things. This means that fast growing markets, where a lot of new consumers are entering the market and new services are introduced daily, can benefit more than markets where the focus is on consolidation and making current services more efficient.  

A valid question in this context is: What is this new generation of cloud services all about?  Mainframes focused on (batch and OLTP) transaction processing, distributed systems were about departmental and integrated planning systems, web systems were about e-commerce (remember when everything was called e-something?) and cloud will be about ...?

My feeling is cloud will be about collaboration, and with collaboration I don't mean offering Twitter or Facebook clones to employees so that they can waste just as much time at work as they do at home :-).  I mean collaboration between multiple parties (organizations, enterprises) in global supply chains. Modern companies will not be vertically integrated like the Ford Motor company was at the beginning of the last century.  They will not do ALL their design, production, distribution, planning, marketing, sales and accounting themselves.  They will consume many or even most of these as services from other service providers. The cloud - Internet-based by design - is ideally suited for that. In my session I went on to discuss the impact of this on traditional IT aspects like security, automation, management etc.

But back to the region... .  The next day I was invited to present a session on cloud computing at the Higher Colleges of Technology in Abu Dhabi.  Abu Dhabi is the largest of the Emirates and the Higher Colleges of Technology are an important source of management talent for the region and work closely with many well-known and esteemed universities and research institutions around the world.  I won't mention the many world leaders and Nobel prize winners that taught or lectured there as I don't want to imply I somehow would fit on that list, but it is clear the region takes technology education very seriously.  And not just for men but also for women -- our session actually took place at the Abu Dhabi Women's College, which features a curriculum stimulating students to start their own companies and, if needed, be able to find work they can do from home.  They are currently planning a new state-of-the-art and very environmentally friendly campus (which is no small feat in a country where the average temperature is above 40 degrees centigrade during the day, and where the need to save energy does not come from a lack of local natural resources.)

What is interesting is the region's natural tendency to institute an ecosystem of service providers.  The Higher Colleges of Technology provide central facilities to about 20 local colleges in addition to an international executive training program.   We see the same in other types of infrastructure -- there is an organization providing IT services to many of the government agencies and companies, such as the local oil and energy companies.  In an economy where investing the proceeds of current economic activities into growth and future activities is the model, the planning of central services and aggregate providers comes natural.  Not that it feels anything like the old archetypes of planned economies with their 5-year plans we know from other regions.  In fact one could argue that using a provider-consumer model seems almost designed to take advantage of the cloud model.

A little example of what I mean by central services.  At some point I managed to lose my suitcase in the trunk of a local taxi.  While my luggage was touring the beautiful city of Dubai, I started a quest to retrieve it.  Turns out the various taxi companies in Dubai (there are about 5, recognizable by different color roofs) all use the same dispatching service, have a joint lost and found department (conveniently located at the airport) and the location of any taxi can be tracked using a combination of GPS and satellite systems.  The only glitch was the fact that a call center was involved (which was as useful and helpful here as call centers anywhere else in the world), but having found a way to circumvent that and talk directly to the central services (by using the personal phone of another taxi driver) I was able to rejoin with my suitcase before leaving later that night.  

As my suitcase now has seen more of Dubai than me, I guess I need to go back soon.  Which is not unimaginable, given the potential cloud seems to have in the region ... despite the fact it only rains about once a year.

Share this post:  
Tags: , , ,
Leave a comment

 

By: CA Community
CA Community is the blog manager’s account used to post general updates and news items.
Read More..

The first 200 servers are the easy part: Private cloud advice and why IT won’t lose jobs to the cloud

Published: October 26 2010, 08:45 AM | no comments
by Jay Fry

The recent CIO.com webcast that featured Bert Armijo of CA Technologies and James Staten of Forrester Research offered some glimpses into the state of private clouds in large enterprises at the moment. I heard both pragmatism and some good, old-fashioned optimism -- even when the topic turned to the impact of cloud computing on IT jobs.

Here are some highlights worth passing on, including a few juicy quotes (always fun):

Cloud has executive fans, and cloud decisions are being made at a relatively high level. In the live polling we did during the webcast, we asked who was likely to be the biggest proponent of cloud computing in attendees' organizations. 53% said it was their CIO or senior IT leadership. 23% said it was the business executives. Forrester's James Staten interpreted this to mean that business folks are demanding answers, often leaning toward the cloud, and the senior IT team is working quickly to bring solutions to the table, often including the cloud as a key piece. I suppose you could add: "whether they wanted to or not."

Forrester's Staten gave a run-down of why many organizations aren't ready for an internal cloud - but gave lots of tips for changing that. If you've read James' paper on the topic of private cloud readiness (reg required), you've heard a lot of these suggestions. There were quite a few new tidbits, however:

==> On creating a private cloud: "It's not as easy as setting up a VMware environment and thinking you're done." Even if this had been anyone's belief at one point, I think the industry has matured enough (as have cloud computing definitions) for it not to be controversial any more. Virtualization is a good step on the way, but isn't the whole enchilada.

==> "Sharing is not something that organizations are good at." James is right on here. I think we all learned this on the playground early in life, but it's still true in IT. IT's silos aren't conducive to sharing things. James went farther, actually, and said, "you're not ready for private cloud if you have separate virtual resource pools for marketing...and HR...and development." Bottom line: the silos have got to go.

==> So what advice did James give for IT organizations to help speed their move to private clouds? One thing they can do is "create a new desired state with separate resources, that way you can start learning from that [cloud environment]." Find a way to deliver a private cloud quickly (I can think of at least one).

==> James also noted that "a private cloud doesn't have to be something you build." You can use a hosted "virtual private cloud" from a service provider like Layered Tech. Bert Armijo, the CA Technologies expert on the webcast, agreed. "Even large customers start with resources in hosting provider data centers." Enterprises with CA 3Tera AppLogic running at their service provider and internally can then move applications to whichever location makes the most sense at a given point in time, said Armijo.

==> What about "cloud-in-a-box" solutions? James was asked for his take. "Cloud-in-a-box is something you should learn from, not take apart," he said. "The degree of completeness varies dramatically. And the way in which it suits your needs will vary dramatically as well."

The biggest cloud skeptics were cited as - no surprise - the security and compliance groups within IT, according to the polling. This continues to be a common theme, but shouldn't be taken as a reason to toss the whole idea of cloud computing out, emphasized Staten. "Everyone loves to hold up the security flag and stop things from happening in the organization." But don't let them. It's too easy to use it as an excuse for not doing something that could be very useful to your organization.

Armijo also listed several tips for finding successful starting points in the move to creating a private cloud. It was all about pragmatic first steps, in Bert's view. "The first 200 servers are the easy part," said Armijo. "Because you can get a 50-server cloud up doesn't mean you have conquered cloud." His suggestions:

- Start where value outweighs the perceived risk of cloud computing for your organization (and it will indeed be different for each organization)
- Find places where you will have quick, repeated application or stack usage
- If you're more on the bleeding edge, set up IT as an internal service provider to the various parts of the business. It's more challenging, for sure, but there are (large) companies doing this today, and it will make profound improvements to IT's service delivery.

Will cloud computing eliminate jobs? A bit of Armijo's optimism was in evidence here: he said, in a word, no. "Every time we hit an efficiency wall, we never lose jobs," he said. "We may reshuffle them. That will be true for clouds as well." He believed more strategic roles will grow out of any changes that come as a result of the impact of cloud on IT.

"IT people are the most creative people on the face of the planet," said Armijo. "Most of us got into IT because we like solving problems. That's what cloud's going to do - it's going to let our creative juices flow."

If you're interested in listening to the whole webcast, which was moderated by Jim Malone, editorial director at IDG, you can sign up here for an on-demand, encore performance.

 

This blog is cross-posted at Data Center Dialog.  Do you tweet?  Follow @jayfry3 on Twitter.

Share this post:  
Tags: , , , ,
Leave a comment

 

By: Jay Fry
Jay Fry is vice president of marketing, Cloud Computing, at CA Technologies. He has over 20 years of experience in marketing and management for innovative enterprise software companies. Prior to CA, Jay was vice president of marketing at cloud computing start-up Cassatt and founded the marketing department...
Read More..

REST and iPhones and Clouds

Published: October 18 2010, 04:35 PM | no comments
by Marvin Waschke

What is REST?

REST (Representational State Transfer) is a distributed application architecture that has been building interest among developers ever since Roy Fielding defined it in his dissertation  in 2000. Fielding was one of the engineers who worked with the World Wide Web Consortium (W3C) and the Internet Engineering Task Force (IETF) to develop the basic designs and standards that made the Internet and the World Wide Web possible (for a brief history lesson on how the Internet and the WWW relate, visit this Wikipedia page). REST is the architecture of the WWW. The ease with which I read the New York Times and my grandsons play online Mario-Brothers-look-alike-games on the same browser is a victory for REST.  The reliability and uniformly good performance of the WWW on the patchwork of carriers and systems that constitutes the Internet is another REST triumph, and let's not forget that the massive scale of search engines like Google all depend on the REST architecture.

REST and the Death of the World Wide Web

With this string of accomplishments, the interest is not surprising. What is surprising, though, is that developers are almost oblivious to these accomplishments. Instead, they are drawn to the simplicity of REST application programming interfaces (APIs). Even more surprising is that the simplicity of REST may be a contributor to the predictions of the death of the World Wide Web last month by Wired Magazine.

Personally, I am cautious about calling REST simple.  I think it does simplify some things, but it complicates others. That's a whole separate discussion.  And declaring the World Wide Web dead is a roaring over-simplification, right up there with declaring IT obsolete. Nothing stays the same, but Amazon has buggy whips in stock.

iPhones

These caveats aside, there is a revealing relationship. The WWW serves up pages described with hypertext markup language (HTML) that a browser elegantly transforms into the palatable eye candy we all enjoy. Apps work differently. Instead of transforming HTML, the apps that are supposedly shutting down the WWW get pure data from the web, process it, perhaps combining data feeds from many remote sources, and do whatever they want with the results: display them to the user, store them, or pass them along to other apps via the web. If an app follows REST architectural constraints, it may be perfectly RESTful.

I don't want to get into the details of what distinguishes a REST API here, but the common alternative to REST is SOAP, which is relatively complex but powerful and flexible way of making remote procedure calls over the web. SOAP requires a substantial stack of software for parsing and managing interactions in addition to the Hypertext Transmission Protocol (HTTP) stack that is already present.

We are now at a point where something a colleague said the other day becomes remarkably insightful:  "Who wants to build a SOAP stack on an iPhone?" This is why REST APIs are a natural for apps and REST APIs are likely to play large in the app revolution when it occurs.

REST and iPhones, and Clouds

What does this discussion have to do with cloud computing?  It relates to a point that James Urquhart made back in 2009 (stay with me, it's still relevant!). He has been asking basic questions for a long time about the role of VMs and operating systems in virtual cloud systems of the future, and has concluded that operating systems are likely to shrink down to a bare minimum, be tailored to the application which will run on them, and almost disappear into the application container. In a way you can think of those sleek and shiny new containers as iPhones, and, as my colleague said "Who wants to build a SOAP stack on an iPhone?"

After writing this post, I decided to take a long vacation, and I am delighted because thinking about REST has exhausted me.  (You have no idea how glad I am to get that off my chest... and pleased to be back.)

*Image used under CCL courtesy of Rob Mercier.

Share this post:  
Tags: , , , ,
Leave a comment

 

By: Marvin Waschke
Marv Waschke is a senior principal architect at CA Technologies. He has represented CA Technologies in several standards groups including the Cloud Management Working Group and Configuration Management Database Federation working groups of the Distributed Management Task Force (DMTF). He is also a member...
Read More..

Delivering a Cloud by Tuesday

Published: October 13 2010, 04:33 PM | no comments
by Jay Fry

Much of what I heard at VMworld in San Francisco (and is likely being repeated for the lucky folks enjoying themselves this week in Copenhagen) was about the long, gradual path to cloud computing. Lauren Horwitz captured this sentiment well in this VMworld Europe preview at SearchServerVirtualization.

And I totally agree: cloud computing is a long-term shift that will require time to absorb and come to terms with, while taking the form of public and private clouds - and even becoming a hybrid mix of the two. How the IT department in the largest organizations will assess and incorporate these new, more dynamic ways of operating IT is still getting sorted out.

Mike Laverick, the virtualization expert and the RTFM blogger quoted in the SearchServerVirtualization story, said that "users shouldn't be scared thinking that they have to deliver a cloud by Tuesday of next week. It's a long-term operational model over a 5- or 10-year period."

But what if you can't wait that long to deliver?

What if you really do have to deliver a cloud (or at least an application) next Tuesday? And then what if that app is expected to handle drastic changes the following Thursday?

Despite the often-repeated "slow and steady" messages about IT infrastructure evolution, I think it's worth making sure we all don't forget that there's another choice, too. After all, cloud computing is all about choices, right? And cloud computing is about using IT to make an immediate impact on your business. While there is a time and place for the slow, steady, incremental change, there's also a very real need now and again to make a big leap forward.

There are times to throw incrementalism out the window. In those situations, you actually do have another choice: getting a private cloud set up in record time with a more turnkey-type approach so you can immediately deliver the application or service you're being asked for.

Turnkey cloud platform trade-offs

A turnkey approach for a cloud platform (which is, in the spirit of full disclosure, the approach that our CA 3Tera AppLogic takes) can get you the speedy delivery times you're looking for. Of course, the key to speed is being able to have all of the complicating factors under your control. The 3Tera product does this, for example, by creating a virtual appliance out of the entire stack: from infrastructure on up to the application, simplifying things immensely by turning normally complicated components into software-only versions that can be easily controlled.

A turnkey cloud is probably best suited for situations where the quick delivery of the application or service is more critical than following existing infrastructure and operations procedures. After all, those procedures are the things that are keeping you from delivering in the first place. So there's your trade-off: you can give yourself the ability to solve some of the messier infrastructure problems by changing some of the rules. And processes. The good news (for CA 3Tera AppLogic customers, anyway) is that even when you do break a few rules there's a nice way to link everything back to your existing environment as that becomes important (and, eventually, required).

Create a real-world example of what your data center is evolving into

For these types of (often greenfield) situations, a turnkey private cloud platform gives you the chance to set up a small private cloud environment that delivers exactly what you're looking for at the moment, which is the critical thing. It also sets you up for dramatic changes in requirements as conditions vary.

But beyond solving the immediate crisis (and preparing you for the next one), there's a hidden benefit that shouldn't be overlooked: you get experience working with your apps and infrastructure in this future, more cloud-like state. You get firsthand knowledge of the type of environment you'd like your entire data center to evolve into over the next 5-10 years. You'll find out the things that work well. You'll find out what the pitfalls are. Most of all, you'll get comfortable with things working differently and learn the implications of that. That's useful for the technology, but also for the previously mentioned processes, and, most of all, for the IT people involved.

So, when is that "big leap" appropriate?

The Forrester research by James Staten that I've referred to in previous posts talks about the different paths for moving toward a cloud-based environment - one more incremental and the other more turnkey. I'm betting you'll know when you're in the scenarios appropriate for each.

Service providers (including some of those working with us that I mentioned in my last post) have been living in this world for a few years now. They are very obviously eager to try new approaches. They are looking for real, game-changing cloud platform solutions upon which they can build their next decade of business. And the competitive pressures on them are enormous.

If you're in an end user IT organization that's facing an "uh oh" moment -- where you know what you want to get done, but you don't really see how you can get there from here - it's probably worth exploring this leap. At least in a small corner of your environment.

After all, Tuesday's just around the corner.

This blog is cross-posted at Data Center Dialog. Do you tweet?  Follow @jayfry3 on Twitter.

Share this post:  
Tags: , , , ,
Leave a comment

 

By: Jay Fry
Jay Fry is vice president of marketing, Cloud Computing, at CA Technologies. He has over 20 years of experience in marketing and management for innovative enterprise software companies. Prior to CA, Jay was vice president of marketing at cloud computing start-up Cassatt and founded the marketing department...
Read More..

A real-time blog from VMworld Europe

Published: October 12 2010, 03:56 AM | no comments
by CA Community

This is a live blog, so appologies apologies  for typo's and possible other errors.

 

This morning 6000 people streamed into the Bella center in Kopenhavn to see the VMworld opening keynote. I'll cover the new items compared to the event in San Francisco. According to VMware's CMO virtualization is about efficiency (utilization), reliability but foremost about Agility. Business' more and more rely or even consist completely of IT and if IT is not agile neither is the business.  Paul Maritz opens by sharing some statistics: the number of X86 deployments on virtual infrastructure has now superseded the deployments on physical infrastructure (now, of course, most of these physical deployments will still exist tomorrow (on their way to becoming legacy) while the virtual ones may be gone tomorrow). More statistics: there are now more than 10 million VMs growing at 28%. There are more copies of Windows and Linux on virtual machines than on physical machines. Maritz called the virtualization layer the "new infrastructure layer," making it very clear VMware is creating the new platform (some might say the new Windows).


According to Maritz, this new infrastructure will be very much about automation and management, with automation going first (management is the necessary evil, the value add comes from automation). This new infrastructure is not just about compute power, but also about storage and networking, but the biggest change will be in security. Security also needs to leave the physical realm and move into the virtual realm with security focused on logical boundaries instead of physical boundaries. Getting security right will be very important to enabling this innovation and agility that virtualization promises. Onto private and hybrid clouds and a new way to "purchase infrastructure". Hybrid clouds is (as described in this blog earlier) defined by VMware as running VMware in your own data center and running VMware in the data center of your service provider(s). Personally I think the world will be a bit more hybrid and so should be hybrid clouds. But it is interesting to see how hybrid is rapidly becoming mainstream (despite apprehensions about security).

Next: Applications. Maritz feels that virtualization is only used to re-host 15 year old (often still batch oriented) applications on a new infrastructure, it will not add the agility the business is looking for. So the new agile world will require new applications build on a new (development) platform.  And - no surprise here -- Maritz sees a big role for Java running on the Springsource framework (again my expectation is that reality may be a bit more hybrid). And just like the virtualization hypervisor layer, these development frameworks isolate the application further form the hardware and make them more portable. By developing in springsource or similar platforms -- like ruby on rails -- application portability again becomes a possibility or even a reality. (Had not heard about this Holy Grail for a while, but do remember -- and have lived through -- the 4GL era, where 4GL tools basically ran your application on anything that came with a plug. Somehow Java killed that movement, but it seems to be making a return as development frameworks that can run on many PaaS platforms (like Google Appworld and Force.com.)

In record speed Martiz went on to Rogue IT a.k.a SaaS applications. He mentioned that 15 applications in VMware somehow made it into the door of VMware and how internal IT is left holding the bag and expected to support them in a compliant and secure manor. Maritz compares this to the entry of rogue PCs and departmental servers entering companies in the '80s. The observation is correct and it is a real challenge. Maritz feelds IT needs to focus on delivering applications and not on managing and monitoring devices (we see the same challenge and feel that adopting a supply chain approach is the pragmatic answer to this, as opposed to a more traditional manufacturing/factory oriented approach).

At this moment Maritz hands over to the CTO to discuss some of the innovations VMware is working on in its product stack, he starts with the core: Vsphere (called "The virtual Giant" on the slide).

Share this post:  
Tags: , , ,
Leave a comment

 

By: CA Community
CA Community is the blog manager’s account used to post general updates and news items.
Read More..

More Posts Next page »