Year: 2010

A couple of weeks ago, during its TechEd Europe 2010 conference, Microsoft announced a new set of programs dubbed Hyper-V Cloud.

The first one is called Hyper-V Cloud Fast Track, developed and executed in partnership with six OEMs: Dell, Fujitsu, Hitachi/HDS, HP, IBM and NEC. It provides reference architectures for private Hyper-V powered Infrastructure-as-a-Service (IaaS) clouds.

In most cases, apparently, there’s no difference between a blueprint for cloud computing and one for virtualization for these vendors.

• The Dell’s one, for instance, has been published in August, and doesn’t mention at all cloud computing.
• Fujitsu just has a landing page and nothing else. It refers to the Fast Track solution as a turnkey solution, but readers can’t do anything except checking the web pages about the products that the OEM will probably include in its reference architecture.
• Hitachi offers a similar landing page, redirecting back to Microsoft website for additional information (which are not available).
• HP, the only quoted in the Microsoft announcement, offers a blueprint called Cloud Foundation for Hyper-V, combining BladeSystem Matrix, System Center products, Hyper-V, and a few consulting services focused on cloud computing.
• IBM suggests an architecture targeting the “midmarket clients” called System x Private Cloud Offering (PCO), which features System x and Brocade networking.
• NEC can just suggest to adopt its Express5800 blade system along with its 10 GB Intelligent Layer 3 Switch, without providing any clear guidance on the storage.

Read more

When in September 2009 the startup tuCloud entered the infrastructure-as-a-service (IaaS) market, virtualization.info reported the news with skepticism. 
The startup was challenging bigger players like IBM in being the first to offer a Desktop-as-a-Service (DaaS) cloud: a form of IaaS where the virtual machines host a client guest operating system rather than a server OS.

A DaaS cloud can also be considered a VDI deployed on a hybrid cloud architecture, where the virtual data center that hosts the virtual desktops is at a cloud provider, while the thin clients are of course at the customer site.
The key problem at that time was that Microsoft licensing terms for VDI, defined in the now defunct Virtual Enterprise Centralized Desktops (VECD) license, were not exactly DaaS-friendly.
tuCloud solved the problem in a simple way: while the company is technically offering DaaS, its cloud is not multi-tenant. This means that portions of the tuCloud cloud infrastructure are exclusively assigned to the customers, and therefore can be considered as extensions of their on-premises data centers. In this scenario, while tuCloud only acts as a Managed Service Provider (MSP) of the cloud infrastructure, a customer can extend his existing Windows Client SA or Volume license.

Read more

November 9 I’ll be at the Cloud Computing Day 2010 in Milan, Italy, giving a lecture about the status and the top challenges of the cloud computing market.

My session will try to clear the air about what cloud computing really is, with a special focus on the Infrastructure-as-a-Service (IaaS) architectures, while discussing the key technology areas where the market still needs to mature. 
I’ll also moderate a panel at the end of the day, igniting the discussion among an interesting group of companies: Acronis, Novell, Red Hat, SonicWALL and Websense.

I’ll be honored to speak along with some top performers like Jan Baan, Founder & Chairman, Chief Innovation Officer at Cordys, and Luisa Bordoni, Country Manager at IDC Italy.

If you plan to attend the event please stop by and say hello: it’s always a great pleasure to meet cloudcomputing.info and virtualization.info readers in person!

Alessandro

Dan Bartow, VP of SOASTA, a load and performance testing company posted a guest blog on the High Scalability website detailing how they used SOASTA CloudTest to simulate 1 million concurrent users on MySpace using 800 Amazon EC2 instances.

In December 2009 MySpace launched a new wave of streaming music offerings in New Zealand, and because of the expected increase of load they wanted to test these features before making them live.

Read more

Last week, at its Symposium, Gartner made an interesting prediction, reported by PCMag: by 2013, only Microsoft and VMware will be perceived as leaders in both cloud computing and enterprise computing.
The authors of this prediction are David Cearley, Vice President, Distinguished Analyst & Gartner Fellow, and David Smith, Vice President of Research Fellow.

Of course the key word is “perceived”, as Gartner analysis is about the customers’ perception on their capability to execute in both private and public cloud computing.
Their analysis included Amazon, Cisco, Google, IBM, Microsoft, Oracle, Salesforce, SAP and VMware.

Cearley said that Microsoft has “one of the most visionary and complete views of the cloud”, despite its offering has yet to mature.
As Amazon and Google are not offering any on-premises solution for enterprises, they aren’t considered as capable. Neither VMware is acting as a public cloud service provider, but Gartner says that it still has a key role as public cloud enabler.

Read more

At the end of last week, Microsoft made an interesting joint announcement with Cloud.com (read virtualization.info coverage of the company): the open source orchestration tool OpenStack, launched by Rackspace and NASA in July, will support Hyper-V as backend hypervisor in the near future, side by side with Xen, KVM and Oracle VirtualBox (plus Citrix XenServer in the near future).
Microsoft won’t develop the code itself, but will work with one of the OpenStack contributors, Cloud.com, to achieve the task.

The announcement has been really vague about how and when, so cloudcomputing.info reached out to Cloud.com to have additional details.

The first thing that is important to clarify is what this integration really means. Cloud.com is developing a hardware abstraction layer for Hyper-V so that OpenStack can use to automate even the most simple tasks, like virtual machines start/stop/resume activities.
This approach will allow customers to use OpenStack with existing Hyper-V virtual data centers without touching the hypervisor.

Read more

It took exactly one year, but Microsoft finally managed to announce how exactly it will extend its existing Platform-as-a-Service (PaaS) public cloud computing platform, Windows Azure, to provide Infrastructure-as-a-Service (IaaS) capabilities too.

A few minutes ago, during its PDC2010 conference, the company announced the Windows Azure Virtual Machine Role for Windows Server 2008 R2. 
To use it, customers will have to build their virtual machines on premises (probably inside a Hyper-V environment as no on-the-fly conversion capability has been mentioned) and then upload them over the Internet to the Azure servers.
Some time in 2011, Microsoft will also allow customers to build new virtual machines directly from inside the Azure control panel, possibly introducing a content catalog and a lab management facility.

At the beginning Azure will only support Windows Server 2008 R2 as guest operating system, but some time during 2011 such support will be extended to Windows Server 2003 and Windows Server 2008 Service Pack 2.

Read more

Despite Microsoft has been really vague about this aspect for months, the market now knows that Windows Azure, its cloud computing platform, will have an Infrastructure-as-a-Service (IaaS) component.
The product is only available as a Platform-as-a-Service (PaaS) public cloud at today, but the company is working with Dell and HP to deliver an on-premises version. And that version will be able to work as a IaaS private cloud too.

What the market doesn’t know yet, is if the public version of Azure will turn into a IaaS+PaaS cloud too. 
Ray Ozzie, the former Chief Software Architect that left the company earlier this week, suggested so in a public statement made one year ago, but nothing else has been said about the topic in the last twelve months.

But now we have an additional hint.

Read more

With a surprising move, yesterday Amazon announced the availability of a free plan for its Infrastructure-as-a-Service (IaaS) cloud computing platform EC2.

The offer, available starting November 1st, will include each month:

• 750 hours of Amazon EC2 Linux Micro Instance usage (613 MB of memory and 32-bit and 64-bit platform support)
• 750 hours of an Elastic Load Balancer plus 15 GB data processing
• 10 GB of Amazon Elastic Block Storage, plus 1 million I/Os, 1 GB of snapshot storage, 10,000 snapshot Get Requests and 1,000 snapshot Put Requests
• 5 GB of Amazon S3 storage, 20,000 Get Requests, and 2,000 Put Requests
• 30 GB per of internet data transfer (15 GB of data transfer “in” and 15 GB of data transfer “out” across all services except Amazon CloudFront)
• 25 Amazon SimpleDB Machine Hours and 1 GB of Storage
• 100,000 Requests of Amazon Simple Queue Service
• 100,000 Requests, 100,000 HTTP notifications and 1,000 email notifications for Amazon Simple Notification Service

Read more

At the beginning of September cloudcomputing.info has been invited by HP to a closed-doors, intimate event for a few influencers in Cupertino. During the two-days mini-conference, very high-profile executives introduced the HP approach to cloud computing, the product and services portfolio, and the go-to-market strategy.

It’s uncommon for a company of this size to pack together in a single room vice presidents, CTOs, and directors to clarify the vision, and answer questions about it for hours, just for a very restricted audience (there were less than 10 people invited, from all around the world). It’s a welcome novel approach considering how fragmented the message can be, buried deep inside tens or even hundreds of press announcements, coming from multiple business units and departments that work on a technology like cloud computing from many different angles.

A few interesting details emerged during the two days, so what follows is a list of things that are worth a report.

Read more