SaaS security provider CloudPassage is filling out its software as a service (SaaS) offering with support for Microsoft's Windows Server platform.
CloudPassage today said its Halo cloud firewall automation and multi-factor authentication software, which customers can provision either in a customer's data center or on a server in a public cloud, now runs on both Linux and Windows Server 2008/R2-based infrastructure. Since launching Halo last year, it only ran on Linux servers.
"When we first started development on the product a couple of years back, Linux was by far the predominant operating system in the cloud," said CloudPassage CEO Carson Sweet. "But since then, there's been a more dramatic shift, where it's gone from 80-20 Linux to Windows, close to 50-50. In some cloud environments we're seeing an even stronger tilt towards Windows. Enterprises are very heterogeneous and they require we support both."
The support for Windows is the latest move by San Francisco-based CloudPassage to try to extend its presence and take on established security suppliers. The company in April received $14 million in Series B financing led by Tenaya Capital joined by earlier investors Benchmark Capital and Musea Ventures, giving CloudPassage a total of $21 million in funding. CloudPassage argues it offers a different approach to securing cloud environments by providing dynamic firewall automation and vulnerability monitoring and assessment.
Halo consists of three core components: The Halo Daemon, a 2MB component that runs on each virtual machine of a cloud server. Each Daemon can run on an infrastructure as a service (IaaS) or platform as a service (PaaS)-based public cloud server or on a private cloud VM. It monitors and reacts to security polices running in the Halo Grid, which as the name implies, is a grid that interacts with each Halo Daemon that provides the core processing. Halo is managed by customers via the Halo Portal.
CloudPassage runs the primary Halo Grid in Rackspace's Cloud Servers, though it is distributed among other cloud providers as well, Sweet said. Customers can run Halo Daemons on other providers' platforms including Amazon Web Services (EC2), GoGrid, Verizon's Terremark and others. It currently is deployed among 17 different cloud providers though Sweet said the company only has relationships with a few of those.
The company was among a number of SaaS providers that last week said it would make its service available on the AppDirect SaaS marketplace. CloudPassage is also running on the Rackspace Cloud Tools Marketplace, also launched last week and powered by AppDirect.
By supporting the Rackspace marketplace, Sweet said customers can procure Halo through Cloud Tools, enabling Rackspace customers to be billed for associated SaaS services on one statement. "Procurement departments will be happy not having to deal with so many small invoices," Sweet said. "They will get one Rackspace invoice."
Posted by Jeffrey Schwartz on 07/25/2012 at 1:14 PM1 comments
The unexpected management shakeup at VMware this week begs the question, is the company succeeding in its effort to reshape its image as a supplier of datacenter virtualization technology to that of a cloud infrastructure and platform provider?
Many also wonder, was the move in which Paul Maritz will be removed as VMware's CEO Sept. 1 to become chief strategist of parent company EMC a reward for a job well done or was he kicked upstairs for not moving fast enough?
More important, is the incoming VMware CEO, Pat Gelsinger, who is currently president of EMC's Information Infrastructure and Products unit (effectively its bread and butter storage business) equipped to meet VMware's goal of "moving to the next phase of cloud computing with the software defined data center and solving the problems of the post PC era," as described by EMC chairman and CEO Joe Tucci, in a conference call announcing the shakeup?
To hear Tucci's reasoning for the changes, he said: "The time to make these kind of changes is from a position of strength when you are performing well and when you have customer permissions to play in these new markets and customers are giving us permission to play in these new markets."
Forrester VP and principal analyst James Staten isn't buying it. "That's good spin but I read it differently," Staten said. "If all was sunny, they wouldn't be unseating the VMware CEO. Our clients tell us they aren't moving up to vCloud (it and they aren't ready). We don't see evidence of success from VMware on the sales of their management tools and they haven't successfully linked CloudFoundry to the hypervisor forming the cloud OS Paul pitched."
Gartner research VP Chris Wolf agreed. "VMware still has its work cut out," Wolf said. "The cloud is extremely competitive and VMware is up against heavyweights such as Amazon, Google, and Microsoft. VMware needs to do a better job incentivizing providers to make VMware software a mainstay in the cloud."
According to a research note this week by Pund-IT principal analyst Charles King in wake of the news , VMware has an edge over its key virtualization infrastructure rivals Citrix and Microsoft.
"VMware has stayed ahead of the game by delivering superior management features and services but most industry watchers believe that the next stage of this market will be defined by increasingly sophisticated hardware/software integration," King noted. "If that's the case, it's difficult to think of a better person to lead VMware than Pat Gelsinger. Along with his deep knowledge of the x86 technologies that are core to VMware's solutions [he was Intel's CTO before arriving at EMC], Gelsinger's intimate understanding of how to get the most out of integrated hardware and software should serve him well in his new position."
Gartner's Wolf agreed but with a caveat. "Gelsinger has a proven track record and should do a good job filling Paul's shoes at VMware," he said. "Pat is not afraid to roll up his sleeves and talk tech, and that will win him favor with VMware's product groups. Still, the pressure will be on. Gelsinger is taking over a company where VMware is the dominant market leader. If Microsoft were to erode a significant part of VMware's market share, then the Gelsinger era would likely be perceived as a failure."
Clearly there's some succession planning going on here as well. Tucci said he plans to stay on at least through the end of next year but it remains to be seen what his plans are afterward, but it is believed he is nearing retirement. While many thought Gelsinger was in line to succeed him, Wolf said EMC may be considering Maritz for the top spot. "Paul is one of the top tech visionaries in the industry, and is well-qualified to lead EMC's strategy in emerging areas such as cloud, big data, and next generation application platforms," Wolf said.
Nevertheless Forrester's Staten has his doubts. It's "not clear [Gelsinger] is the right guy to replace Paul or the guy VMware needs right now but he was next in line," he said. "Nothing in his background suggests he is better equipped. We'll see if he can prove there's more to this than pure succession and EMC tightening the reins on VMware."
Posted by Jeffrey Schwartz on 07/19/2012 at 1:14 PM0 comments
Rackspace is making it easier for enterprises using its cloud offerings to procure and integrate software as a service (SaaS) tools and apps.
The hosting provider this week launched its Cloud Tools Marketplace, which aims to let its customers find and ultimately subscribe to any number of SaaS offerings. Cloud Tools is built on the marketplace platform of closely-held AppDirect. Rackspace describes the new marketplace as an upgrade, though until now the company effectively offered static listings of partners. Cloud Tools lets customers view demos such as videos and screen shots and allows them to subscribe to the services.
Among those on Cloud Tools are cloud management providers enStratus, RightScale, and ScaleXtreme; monitoring and reporting tools vendors Cloudability and New Relic; app deployment provider Standing Cloud; SaaS security operators CloudPassage, Dome9 and Still Secure; and E-mail provider SendGrid.
Rackspace's new marketplace lets customers tie these and other various SaaS offerings to their accounts and receive one bill for the use of those collective services, removing the need to subscribe individually to those services directly from the provider. Ven Shanmugam, Rackspace senior manager of corporate strategy says Cloud Tools will provide a more "seamless experience" for customers using its infrastructure as a service (IaaS) cloud offerings.
Cloud Tools also raises the visibility of AppDirect. Launched in 2009, AppDirect's marketplace lets providers of all types, ranging from telecom, system integrators and even retailers tie SaaS offerings to their respective service offerings. On top of this week's announcement that it is hosting Rackspace Cloud Tools (which will run on Rackspace Cloud), AppDirect disclosed an $8.5 million round of venture financing led by iNova Capital, which adds to last year's $3.25 million round.
While Rackspace is among the most prominent cloud provider AppDirect has signed on, it is also running marketplaces for mobile developer platform provider Appcelerator, Deutsch Telecom and Bell Canada. The company will be shortly announcing a major retail win as well, said AppDirect's co-founder and CEO Daniel Saks. Unlike app stores hosted by say Apple or even the forthcoming Microsoft Windows Store, Saks said his company's marketplace platform is suited to those looking to procure subscription-based SaaS services.
The AppDirect marketplace platform is built on a Java-based stack, supports RESTful APIs and enables authentication and authorization using OAuth and OpenID.
While AppDirect doesn't offer its own marketplace, it has partnerships with numerous SaaS providers such Docusign and Google. Those using the AppDirect platform can leverage those partnerships by offering them on their marketplaces, Saks explained. "Our vision was to create the fabric for the cloud that connects the developers of great applications to businesses through a network of app stores," he said.
Posted by Jeffrey Schwartz on 07/18/2012 at 1:14 PM0 comments
Looking to include the fees from using public cloud services into your IT budgets? Cloud management provider RightScale today launched a free Web-based service that helps customers predict cloud costs.
RightScale, one of the largest independent cloud management providers, said it has acquired ShopForCloud for an undisclosed amount and re-launched the service as PlanForCloud.
The cloud forecasting service is designed to help customers predict the cost of using public clouds based on various usage models including compute, storage, file transfer and database. The models take into account usage patterns and fluctuating prices from different cloud providers. Based on those models, the service creates forecasts and generates reports.
"Knowing the cost of cloud services is critical whether companies are making their first foray into cloud computing, planning to deploy additional resources or managing their overall budget for cloud spending," said RightScale CEO Michael Crandell, in a statement. "PlanForCloud provides users a forward-looking view of costs and perfectly complements our existing cost reporting and tracking features in the RightScale Cloud Management Platform. The combination will allow businesses to manage their current and past cloud spending as well as their future cloud budgets."
PlanForCloud pulls configuration and pricing data from Amazon Web Services, Google Compute Engine (GCE), Microsoft Windows Azure and Rackspace. RightScale said it plans to work with other cloud providers in the future. The service lets customers forecast costs of the cloud service providers supported and creates three-year reports to help IT managers incorporate the projected cost of those services into their budgets.
RightScale is offering the free service to anyone, whether or not they are existing customers. Naturally the company hopes the service will be a draw for its cloud configuration, provisioning, automation, monitoring and governance services, which provide common management of private, hybrid and multiple public cloud services.
Posted by Jeffrey Schwartz on 07/18/2012 at 1:14 PM0 comments
Coming off last month's announcement that its Windows Azure service will let customers run virtual machine instances, Microsoft this week took a step toward enabling third party hosting providers to build cloud service offerings running on Windows Server.
Microsoft released the community technology preview (CTP) of Windows Server for Hosting Providers. Based on Microsoft's forthcoming Windows Server 2012, the suite will allow hosting providers to provision and manage through System Center, Windows Azure Web Sites, virtual machines, the Service Management Portal and API.
The move should aid hosting providers that want to offer services that are compatible with Windows Azure. Microsoft announced the CTP release at its Worldwide Partner Conference, which took place this week in Toronto.
Indeed, hosting providers are gunning to offer services that are compatible with Windows Azure and private clouds running Windows Server, said Sinclair Shuller, co-founder and CEO of Apprenda, which last month launched Apprenda Azure.
As the name implies, Apprenda's software will let organizations create private clouds that have the same characteristics as Windows Azure service except they run in a customer's datacenter. Nevertheless customers can also shift workloads or create hybrid clouds that split workloads, applications and data between Windows Azure and Apprenda Azure, Sinclair said.
In concert with Microsoft's announcement this week, Apprenda joined three other vendors -- Cloud Cruiser, Derdack and ServiceMesh -- to create the Private Cloud Solutions Suite (PCSS). All four are Microsoft ISV partners. Cloud Cruiser offers a billing and chargeback system, Derdack's Enterprise Alert provides alerting and notification and the ServiceMesh Agility Platform is used for provisioning, governance and management of hybrid cloud applications.
"Microsoft doesn't have all the moving parts to provide a full robust full blown private cloud deployment platform," Sinclair said. Maybe not now but what will Sinclair do if Microsoft decides to offer one in the future?
"They could do whatever they want but in the context of private cloud but we don't see Azure coming on premises," he said. "They've talked about it in the past but it hasn't materialized. We work very closely with the team and we don't have any personal fear that's going to happen. Microsoft wouldn't be telling 1,000 systems integrators worldwide they should be looking at Apprenda for private PaaS if they planned on jumping in."
Posted by Jeffrey Schwartz on 07/12/2012 at 1:14 PM0 comments
Google made it official this week that it will offer an infrastructure as a service cloud, dangling performance as its value.
The launch of Google Compute Engine at the company's annual Google I/O developer conference was widely expected, though overshadowed by its foray into the tablet market with its own Nexus 7. By launching an IaaS, Google is taking on some large players including Amazon Web Services, Rackspace, Verizon, IBM, Hewlett-Packard and Microsoft, which earlier this month launched its long-anticipated IaaS services.
Like Microsoft, whose Windows Azure was just a platform as a service, Google's cloud portfolio other than its Google Apps service, was the Google App Engine PaaS. But in launching GCE, which is now in limited preview, Google is taking it slow. For now it's only offering Linux virtual machines (translate: no Windows) support. While there's plenty of demand for Linux, Google is still only reaching out to a limited segment of the enterprise market.
And will it play ball in supporting any of the other cloud vendors' platforms and APIs to enable portability with public, private and hybrid clouds? The IaaS market is fragmented with a variety of compute platforms such as OpenStack, CloudStack and of course Amazon.While these and other players such as Eucalyptus and Nimbula have services that let companies move instances from Amazon to other public and private clouds (as does Citrix CloudStack), we are a long way from removing true cloud lock-in as a barrier.
And vendor lock-in, according to a Rackspace-commissioned study released this week, is a key consideration by 86 percent of customers in choosing a cloud vendor. While Google has yet to make a strong public proclamation on where it stands with regard to IaaS portability (that is, will it join or support OpenStack or CloudStack efforts?) there are some aspects of GCE worth noting:
- Google appears to be targeting compute-intensive workloads, particulary for compute-intensive applications requiring 100 virtual machines or more.
- Customers can store their data locally, on a new persistent block storage device it is offering or via its existing Google Cloud Storage service.
- GCE will offer networking capabilities to enable customers to create and manage their compute clusters.
- Developers can either configure and control their VMs using a scriptable command-line tool or Web UI or use Google's APIs to build or link to a management system. Google has partnered with RightScale Puppet Labs, OpsCode, Numerate, Cligr and MapR, whose tools will integrate with GCE.
Craig McLuckie, the product manager for GCE said in a blog post that Google intends to use its vast infrastructure to power GCE. "This goes beyond just giving you greater flexibility and control," he noted. "Access to computing resources at this scale can fundamentally change the way you think about tackling a problem."
Do you think Google will emerge as a major IaaS player? What does it need to get there? Drop me a line at firstname.lastname@example.org.
Posted by Jeffrey Schwartz on 06/29/2012 at 1:14 PM1 comments
It's certainly not a shocking revelation that enterprise customers don't want to be tied to any one cloud service, but a new survey commissioned by Rackspace shows 86 percent see portability as a key consideration when choosing a provider.
Rackspace CTO John Engates said in an interview this week that concern over provider lock-in validates its decision to spawn the OpenStack initiative, borne nearly two years ago with the contribution of cloud compute code co-developed with NASA. Now with 180 software, hardware and infrastructure providers on board, Engates said OpenStack should take vendor lock-in off the table.
"OpenStack clouds have a better story because the more fragmentation that exists between cloud providers, the more people will gravitate to something that looks and feels and starts to have familiarity between different cloud providers," Engates argued.
But will they really gravitate to OpenStack to get that portability? Consider Engates' recent remarks targeted at Citrix, in which he called out the vendor for pulling away from its earlier support of the OpenStack compute initiative in favor of its own compute engine and ecosystem CloudStack. There was clearly bad blood when Citrix, an early OpenStack participant and proponent, jumped off the bandwagon and released to the Apache Software Foundation CloudStack as an alternative open-source cloud compute project it acquired last year.
Engates said in a keynote address at the Open Data Center Alliance Forecast 2012 conference two weeks ago that Citrix should have combined CloudStack with OpenStack's Nova compute engine. In response, Sameer Dholakia, group VP and general manager of Citrix cloud platforms group, said at Cloud Expo in New York Citrix decided to decamp from OpenStack and push forward with CloudStack after six months of unsuccessfully trying to bring it together with Nova.
Jonathan Bryce, chairman of the OpenStack Project Policy Board and co-founder Rackspace Cloud responded to those charges saying Citrix never made any moves to submit CloudStack to OpenStack.
"We have these public processes for submitting new projects for incubation and potential inclusion and they never formally submitted any of these things," Bryce said. "They had discussions with different community members but I don't think they submitted blueprints or an incubation proposal to include their CloudStack software into OpenStack."
Combining the two, he added, would be technically challenging. "OpenStack is a bunch of Python services that tie into all these different technologies and CloudStack is basically a Java application," he said. "There was already a different approach from the technology standpoint and how they were built."
Bryce said while that will result in two open source compute cloud systems, he didn't see that as a major setback for OpenStack. "It's not a huge disaster," he said. "It's a very big market and they're different companies with different needs and different types of clouds. What we have tried to build with OpenStack is flexible and scalable cloud components that include compute but also advanced networking and storage as well. Just looking at the breadth of the technology included in OpenStack, and the size of the community we have with the number of developers and the number of companies we have including Cisco, Dell, HP, IBM and Rackspace. We have a pretty broad base of support and that's why we're still moving forward."
But there are still other camps. Among others, there's Microsoft's Windows Azure, which is now emerging into an IaaS, the newly announced Google Compute Engine and of course the huge behemoth Amazon Web Services, which is believed to be the most widely used IaaS. Many players are looking to forge interoperability with Amazon, including CloudStack, Nimbula (a startup formed by the early developers of Amazon Web Services) and Eucalyptus, which recently announced an agreement with Amazon to allow Eucalyptus customers use of Amazon's APIs.
Rackspace's Engates was unbowed by the Eucalyptus-Amazon API sharing pact, which lets customers move Amazon workloads to public and private clouds running the Eucalyptus platform.
"You can try to mimic Amazon APIs, but you are going to be in this constant cycle of reverse engineering whatever Amazon has done and trying to copycat the technology underneath," Engates said. "You don't have Amazon's technology, you don't have their underpinning. You're really trying to build a surface level compatibility without the underlying technology and that's very hard to do in a sustained way."
Eucalyptus CEO Marten Mickos said in an e-mail that it has a large stable of customers already using clouds that are compatible with Amazon. "It's the customers that matter," he noted. "There is a rapidly growing number of them. They love our product, they run hybrid clouds, they are experiencing the everyday benefits of API compatibility. Any statement to the contrary is ludicrous and says more about the one who uttered it than about APIs."
Asked about the implications of OpenStack, he replied: "We will welcome OpenStack as a true public cloud service for the world. If and when they reach a sizeable market share, we will be happy to support whatever API they at that point operate with."
So while users of cloud services are concerned about vendor lock-in, there are a number of viable options for certain levels of portability. OpenStack is clearly one of them with substantial and prominent backers but it's not the only game in town.
Posted by Jeffrey Schwartz on 06/29/2012 at 1:14 PM0 comments
Over the past year, Salesforce.com CEO Marc Benioff has evangelized the social enterprise and the shift by workers to mobile devices. In addition to enabling social networking by bringing out apps such as Chatter and acquiring companies like Radian 6 last year and Buddy Media earlier this month, the company has its cloud-based platform, Force.com and Database.com, that lets developers and ISVs integrate their apps with Saleforce.com.
Just how formidable is Force.com these days? Company officials say the platform recorded 54 billion transactions last quarter, hosts 300,000 custom apps and its AppExchange, launched in 2005, hosts 1,500 ISV applications.
With its emphasis on the social enterprise and mobile users, Salesforce has extended its platform to ease the development of next-generation apps. The company updates the platform three times a year and today has issued its latest release, called "Summer 12," which officials said will make it easier for developers to add social networking capabilities to legacy and new apps and target them to mobile devices.
Targeting mobile clients is hardly unique these days. Forrester Research estimates one-fifth of those with mobile devices use them to go online and mobile users will outnumber people using their PCs within four years.
The Summer 12 release gains native geo-location fields in database.com for location-aware apps, typing them to maps. "Every object in the database first gets geo-located, you can imagine the pretty amazing set of apps that are going to come from that," said Adam Seligman, Saleforce.com's VP of developer relations.
Salesforce has also added a new Web-based tool called Schema Builder that lets developers drag and drop cloud databases with via REST APIs within database.com. Additional REST-based APIs included to target mobile devices accessing Chatter that lets developers collaborate on any object within database.com.
"The cycle of trying to get a database provisioned and to define what your schema ought to be, while making sure the security model is going to work and be consistent across your organization all gets simplified down to one cloud database," Seligman said. To improve security, the platform now has support for OAuth, a standard for authentication.
Seligman said the new platform makes it easier for developers to embed security controls into their apps. Connected Apps lets developers include pin-level security. The new APIs support streaming enabling developers to allow their apps to be updated in real-time, while using the security controls, he explained.
All of this is done as an alternative to messaging middleware and traditional application servers, added Quinton Wall, director of technical platform marketing at Salesforec.com. The ability to push updates to applications out in real time means ISVs and partners can release updates when they choose, Wall told me (he elaborated on Summer 12 in a blog post). "Everything is becoming real time in the enterprise," he said. "We expect that in our lives and developers expect that as well."
Posted by Jeffrey Schwartz on 06/19/2012 at 1:14 PM0 comments
Riverbed is best known for its WAN acceleration appliances but the company also offers software designed to improve Web app performance in virtual and cloud environments. Now Riverbed is tying its software that optimizes Web site performance with its application delivery controller (ADC), designed to automate load balancing in order to provide app scalability.
The company's new Stingray Traffic Manager 9.0, announced yesterday, combines the ADC with its Aptimizer Web content optimization software. Naveen Prabhu, a senior product marketing manager at Riverbed pointed to research by various players including Forrester Research, Google and Microsoft that concluded that anytime a user has to spend more than 250 milliseconds more on one site than another. Since many application environments already have ADCs, Riverbed sees adding Web site optimization to them as a natural extension.
This software offering is typically found in virtual data centers and cloud environments, according to Prabhu. "This is the only ADC or load balancer that's purpose built for the cloud because it's designed to take advantage of your off the shelf hardware and you can add more performance just by throwing more hardware at it," he said. Stingray is used by a number of prominent cloud providers including Amazon Web Services, Joyent, RightScale and Rackspace, Prabhu said.
Until now, Riverbed offered the optimization technology as a plug-in to Windows IIS and SharePoint servers but now the company is integrating it into the Stingray ADC (as a license upgrade). Riverbed is also now offering the optimization software as a VM for virtual and cloud environments.
The new software costs $18,500 and covers up to 250,000 page views. For those requiring more capacity the cost is $7,500 for each additional 250,000 page views.
Posted by Jeffrey Schwartz on 06/19/2012 at 1:14 PM0 comments
Those that rely on Amazon Web Services for their compute infrastructure felt the ramifications of that dependence last week when a power outage brought on by a cable fault in its North Virginia Availability Zone shut down a number of sites including Quora, Pinterest and Salesforce.com's Heroku platform as a service cloud.
The overnight outage lasted a few hours but it gave proponents of private and hybrid clouds cause to extol the virtues of distributing infrastructure and applications to internal datacenters and/or collocation facilities. Coincidentally, a number of software vendors this week are rolling out wares designed to give enterprises the benefits of public cloud computing within their own datacenters. ActiveState, Eucalyptus and Virtustream introduced new software aimed easing the sharing of compute and application workloads between public and private clouds.
While these announcements were planned in advance of Amazon's latest outage at least one seized the opportunity to call attention to itself in wake of the disruption. ActiveState, a company that offers software for running applications on platform as a service (PaaS) clouds, reached out yesterday on the premise of the Heroku outage. "We're being a little opportunistic, to be honest," admitted Toph Whitmore, ActiveState's VP of marketing. "But let me be clear, Heroku's disruption from Amazon's outage is not something to gloat about, it hurts everybody. If anything it could really further delay consideration for cloud computing adoption whether it's public cloud or private cloud."
The new 1.2 version of ActiveState's Stackato software allows developers host PaaS-style clouds in their own datacenters. The new release allows for the migration Heroku Buildpacks from the public cloud to datacenters running Stackato. As a result, customers can move applications built to run on Heroku, which was unavailable during the Amazon outage, in their datacenters without modifying any of the apps underlying code.
Heroku Buildpacks are scripts that are used to build and host specific application frameworks. "Never say 'never' but there is no recoding necessary for Heroku apps to run on Stackato," Whitmore said. "Now there is a migration path for enterprises running on Heroku who need something a little more secure."
For those looking to migrate their Amazon workloads, there are a number of options. As I reported last week, Citrix is positioning CloudStack as a compute infrastructure suited for the development of private and public clouds that support Amazon's APIs. AppFog, Nimbula and Eucalyptus also offer software offers compatibility with Amazon (AppFog's PaaS software also supports HP Cloud, OpenStack and Windows Azure).
Eucalyptus today launched version 3.1 of its namesake platform. The new release brings together the free open source version and the commercially supported distribution under one platform, said Marten Mickos, Eucalyptus CEO. "It is the same underlying platform, there's just one development tree," Mickos said. A key new feature in the new release, called FastStart, provides a self-service, automated setup procedure for building infrastructure as a service (IaaS) clouds.
The FastStart capability, which Mikos said allows a user to set up a cloud in 20 minutes, can be set up on CentOS 5 with the Xen hypervisor or CentOS 6 with KVM. FastStart provides access to what Eucalyptus calls euStores, pre-built cloud images. For those that want to build more sophisticated images, the company offers SilverEye, consisting of installation and configuration tools.
For those not requiring FastStart and looking for enterprise grade deployments, Eucalyptus 3.1 now runs on Red Hat Enterprise Linux for Amazon EC2, EBS S3 storage and Identity and Access Management (IAM) running VMs from Red Hat or VMware. Also starting with the new release, Eucalyptus is making all source code available on GitHub, where all development activity will take place and defect and feature tracking will be posted.
Virtustream, which operates six datacenters worldwide, today said it is offering the software that powers its cloud infrastructure to customers who want to shift their workloads between public and private clouds. Chief marketing officer Simon Aspinall said Virtustream is moving to compete with companies like Eucalyptus and Nimbula but argues its new xStream 2.0 software offers "a private cloud tailored to the enterprise, high security, high compliance, high performance guaranteed model."
Aspinall said the software offers silicon-level authentication with Intel TxT and supports a variety of security and compliance standard including FISMA, HIPAA, PCI and the SAS 70 and SSAE 16 audit standards, among others. The software can meet service level agreements by dynamically integrating compute, storage and networking into what the company calls MicroVMs, which provides multitenant virtualization.
Virtustream's software also offers the ability for service providers to offer consumption-based pricing and for enterprises to establish chargeback based on usage in increments as little as 5 minutes. The software is entering beta this week and will be generally available next month, Aspinall said.
For those with compliance or governance restrictions that prohibit them from putting certain types of data in the public cloud, these new options, among others, promise to make it more practical to implement hybrid clouds, while providing platforms for moving apps and workloads between public and private clouds.
Posted by Jeffrey Schwartz on 06/19/2012 at 1:14 PM0 comments
There is some really bad blood between executives at Rackspace and Citrix and the result is two fragmented open source Infrastructure as a Service (IaaS) compute platform efforts. Perhaps it's not surprising that Rackspace felt jilted by Citrix's decision to abandon its commitment to the Rackspace-founded OpenStack project in favor of its newly acquired CloudStack platform, but the fur was flying at this week's Cloud Expo and the Open Data Center Alliance (ODCA) Forecast 2012 conference in New York.
Rackspace CTO John Engates, who is the company's cloud computing evangelist, aired his disdain for Citrix's move (which was underway for some time but became official in April) during a keynote address at the ODCA conference Tuesday. "I don't know what Citrix was thinking when they moved away from OpenStack because OpenStack had all the momentum in the world," Engates said in response to a question about the matter.
The OpenStack Project, developed by Rackspace and NASA (the latter of which apparently is now moving some of the space agency's cloud compute to Amazon Web Services), kicked off nearly two years ago and has since been backed by 180 companies including AT&T, Cisco, Dell, Hewlett-Packard, IBM, Intel and Red Hat. HP in particular is basing its new HP Cloud service on OpenStack. Despite the momentum of OpenStack, critics complained that the platform has evolved slowly, as has the open source governance model.
Yet Engates suggested that Citrix made a short-term bet that acquiring Cloud.com last year and deciding to base its cloud-compute architecture on CloudStack rather than stick with its earlier plan to use the OpenStack Nova compute engine would be a quicker path to revenue, given the evolution of the two.
"I think, and this is just my opinion, they didn't make the right choice," Engates said. "They should have stuck with us, they should have combined efforts instead of fragmenting. They did make the choice that they made and we'll see what happens."
In a keynote address at Cloud Expo on Wednesday, Sameer Dholakia, group VP and general manager of Citrix cloud platforms group, acknowledged the rumblings from the OpenStack camp and explained why Citrix decided to decamp from OpenStack and push forward with CloudStack.
"We did an assessment a little over a year ago, that where we were with the OpenStack project was at least a year-and-a-half away from being ready for prime time and in production," he said. "And we needed to be in market providing our customers answers and competing with other vendors in the market like VMware that had vCloud out there. And so we decided that we had to move faster than the community could create. So we made the acquisition of cloud.com."
In response to Engates' charges that they should have combined efforts, Dholakia said they tried for over six months to find a way to bring CloudStack and the Nova compute engine of OpenStack together.
"We had the sincerest of efforts to try to bring these two things together," he said. "We simply weren't able to get there. I wish we were able to get there but we weren't. It certainly wasn't for a lack of trying. But I will tell you that the cloud really is now. People are building clouds today. They need software with which they could go build those real clouds and put into production today."
Dholakia said there are more than 100 clouds powered by CloudStack including those run by telecom giant BT, GoDaddy, Zynga and Samsung. In an effort to illustrate why OpenStack wasn't ready for prime time and CloudStack was, Dholakia said that Samsung had spent more than a year trying to build a cloud infrastructure on OpenStack to support a service for its chat and messaging services. Samsung had "an army of developers" that were unable to get OpenStack into production. "They called us and within two weeks had a proof of concept running and it's now in production and deployed," Dholakia explained.
Citrix's decision to contribute CloudStack to the Apache Software Foundation has further boosted its cache. Another factor that gives it appeal is the fact that it supports Amazon's APIs. Indeed, in conversations with various stakeholders at Cloud Expo this week, the sentiment generally was that the CloudStack compute platform was more ready for prime time than OpenStack, but the latter had more under its belt with the Swift block storage architecture. Moreover, the work on OpenStack Quantum, the networking project in which Cisco is lending a major hand, also has a lot of promise.
Yet many that have initially backed OpenStack have jumped on the CloudStack bandwagon. One such example is Dallas-based SoftLayer, a cloud provider with 13 datacenters worldwide, 100,000 servers and $400 million in revenues, which this week launched a private cloud service that will be based on CloudStack.
SoftLayer CTO Duke Skarda told me when the company decided to build its private cloud service it considered Eucalyptus, CloudStack, OpenStack, VMware's vSphere, Nimbula and Microsoft. In the end it came down to CloudStack and Eucalyptus, which also now supports Amazon's APIs.
"Ultimately we went with CloudStack. Of all of those it's probably the most mature," Skarda said. "CloudStack has the maturity, a suite of tools and great support behind it in Citrix. that's the reason we've deployed them. Using Citrix's CloudBridge, it could extend clouds in customers' datacenters to its own," he added.
Also an OpenStack backer, Skarda said it is using the OpenStack Swift object store for storage. Asked how he sees the feud between the two compute camps playing out, Skarda said it looks like there will be fragmentation for some time to come. "Right now it doesn't look like there's a path for them to come together," he said. "Over time, the fact that CloudStack is part of Apache certainly lends it a lot of credibility, but OpenStack has a huge momentum behind it. But right now it seems like there's a fragmented market coming."
Certainly, this is not the first time we've seen fragmented computing efforts. Remember the Unix wars with Solaris, AIX, HP/UX and other instances? And of course, as Linux matured, many wondered if having multiple distributions would lead to fragmentation. It certainly hasn't been a cakewalk, adding to the mix other broadly deployed platforms, notably Windows.
Do you think OpenStack and CloudStack need to come together? Or will they and all the various other alternatives offer customers the options they need to push forward with their cloud computing initiatives?
Drop me a line at email@example.com or leave a comment below.
Posted by Jeffrey Schwartz on 06/15/2012 at 1:14 PM2 comments
While Amazon Web Services is the undisputed leading provider of cloud computing services, competitors often are able to legitimately put forth a value proposition centered on customer support. That argument might become a tad more difficult now that Amazon has said it will offer higher-touch customer support.
Amazon announced today it is extending its free support and reducing the cost of premium tiers while adding more support services. The company also said it is will be more proactive in alerting enterprise customers of opportunities to lower their costs or improve performance.
Under the new plan, Amazon is revising its four tiers: Basic, Developer (previously Bronze), Business (previously Gold) and Enterprise (previously Platinum). Here's a rundown of the new tiers as described by the company:
- Basic: Newly added features to its free support include 24-hour access to customer service via e-mail or phone for technical support, billing and account questions.
- Developer: Priced at $49 per month, this service guarantees 12-hour response time and access to Amazon support engineers via e-mail during local business hours for help configuring and operating Amazon's core services.
- Business: At this level, customers will receive responses within one hour from an "AWS Trusted Advisor." This tier provides 24x7 phone, e-mail and chat support. Also covered is support for common third-party software (such as operating systems, databases, etc.). Amazon slashed the price of this service from $400 to $100 per month. Plus, it lowered usage fees from 5 to 10 percent to 3 percent.
- Enterprise: Customers running mission-critical apps receive 15-minute response times plus a dedicated technical account manager for various planning and reporting services. Amazon has reduced the minimum fee from 10 percent of usage to variable costs ranging from 3 to 10 percent.
Amazon will also offer what it is calling "trusted advisor updates," which is based on best practices from serving customers. "It inspects your AWS [Amazon Web Services] environment and makes recommendations when opportunities exist to save money, improve system performance or fault tolerance or close security gaps," said Amazon Web Services technical evangelist Jeff Barr in a blog post.
For example, Barr said it checks Elastic Load Balancer (ELB) optimization, recent Elastic Block Storage (EBS) warnings, EC2 Availability Zone balance and service limit checks to ensure a customer is not nearing limits on instances, volumes or IP addresses. He also pointed to a new self-service portal that lets customers view recommendations and observations covering such areas as security, fault tolerance and cost optimization.
Posted by Jeffrey Schwartz on 06/14/2012 at 1:14 PM0 comments