Category Archives: Infrastructure as a Service

Spring.io: Containers are great, but need to be simple to adopt

Spring.io offers containers as a service

Containers are all the rage but everyone does them differently

ElasticHosts founder Richard Davies launched another container-focused venture this week – Spring.io, a pay-as-you-use cloud service targeted primarily at Linux developers. Davies told BCN that while the volume around Docker and other container technologies is high, which is encouraging, their uptake will ultimately depend on how providers balance simplicity with performance.

Spring.io is a spinoff of Davies’ other venture, ElasticHosts, which still uses Linux containers but operates more like a traditional infrastructure as a service provider (customers need to subscribe to the service for set periods).

“We have been listening to the market and what we are hearing is that people are craving simplicity,” Davies said. “They just want to be able to sign up to a service without having to choose instance sizes or worry about over-paying, just as you would with your gas or electricity.”

The benefit containers offer over traditional virtualisation platforms is that they scale more closely in line with the resources an application needs, and they scale much more quickly. But most cloud services are provisioned in fixed virtual and / or physical increments that can only scale by adding or subtracting fixed-size VMs or hardware or both. This, Davies said, leads to massive amounts of waste in terms of asset utilisation (for the provider) and cost (for the consumer). It’s a ‘lost-lose’.

Spring.io uses the same underlying technologies as Docker and other Linux containers (cgroups, namespaces), but whereas they are mostly application containers with a focus on simplifying portability and managing dependencies in micro-services architecture, Springs.io offers operating system containers with a focus on usage-based billing and reactive auto-scaling.

Where most container as a service providers lean heavily upon scripting for scaling and deployment , or run containers within virtual machines, Springs.io supplies auto-scaling for load straight out of the box and requires no user API calls or JSON-juggling for their management or monitoring.

“Docker is exciting to developers struggling with shipping applications, we believe Springs.io is exciting to devops and sysadmins that want simple scaling for a reasonable price,” Davies.

HP, CenturyLink buddy-up on hybrid cloud

CenturyLink and HP are partnering on hybrid cloud

CenturyLink and HP are partnering on hybrid cloud

HP and CenturyLink announced a deal this week that will see HP resell CenturyLink’s cloud services to its partners as part of the HP PartnerOne programme.

As part of the deal HP customers will have access to the full range of CenturyLink services, which are built using HP technology, including managed hosting, colocation, storage, big data and cloud.

“CenturyLink solutions, powered by HP, provide compelling value for organizations seeking hybrid IT solutions,” said James Parker, senior vice president, partner, financial and international, at CenturyLink. “CenturyLink complements the HP portfolio with a breadth of hybrid solutions for enterprises, offering customers the ability to choose the services that make the most sense today, while retaining the flexibility to evolve as business demands shift.”

HP said the move will help CenturyLink expand its reach new customers, with HP exploiting new opportunities to build hybrid cloud solutions for existing customers.

“As businesses map out a path to the cloud, they need flexibility in how they consume and leverage IT services,” said Eric Koach, vice president of sales, Enterprise Group, central region, HP.

“HP cloud, software and infrastructure solutions help CenturyLink and HP enable clients to build, manage and secure a cloud environment aligned with their strategy, across infrastructure, information and critical applications,” Koach said.

Since splitting up HP has bifurcated its partner programmes into the PartnerOne programme for service providers and the Helion PartnerOne programme, the latter of which largely includes services providers building solutions on top of OpenStack or Cloud Foundry.

DataCentred ARM-based OpenStack cloud goes GA

DataCentred is moving its ARM-based OpenStack cloud into GA

DataCentred is moving its ARM-based OpenStack cloud into GA

It has been a big week for ARM in the cloud, with Manchester-based cloud services provider DataCentred announcing that its ARM AArch64-based OpenStack public cloud platform is moving into general availability. The move comes just days after OVH announced it would roll out an ARM-based cloud platform.

The company is running the platform on HP M400 ARM servers, and offering customers access to Intel and ARM architectures alongside one another within an OpenStack environment.

The platform, a product of its partnership with Codethink originally launched in March, comes in response to increasing demand for ARM-based workload support in the cloud according to DataCentre’s head of cloud services Mark Jarvis.

“The flexibility of OpenStack’s architecture has allowed us to make the integration with ARM seamless. When users request an ARM based OS image, it gets scheduled onto an ARM node and aside from this the experience is identical to requesting x86 resources.  Our early adopters have provided invaluable testing and feedback helping us to get to point where we’re confident about stability and support,” Jarvis explained.

“The platform is attracting businesses who are interested in taking advantage of the cost savings the lower-power chips offer as well as developers who are targeting ARM platforms. Developers are particularly interested because virtualised ARM is an incredibly cost-effective alternative to deploying physical ARM hardware on every developer’s desk,” he added.

The company said ARM architecture also offers environmental and space-saving benefits because they can be deployed in higher density and require less power than more conventional x86 chips to run.

Mike Kelly, founder and chief executive of DataCentred didn’t comment on customer numbers or revenue figures but stressed the move demonstrates the company has successfully commercialised OpenStack on ARM.

“The market currently lacks easy to use 64-bit ARM hardware and DataCentred’s innovation provides customers with large scale workloads across many cores. Open source software is the future of computing and the General Availability of DataCentred’s new development will make our services even more attractive to price-sensitive and environmentally-aware consumers,” Kelly said.

DataCentred isn’t alone in the belief that ARM has a strong future in the cloud. The move comes the same week French cloud and hosting provider OVH announced plans to add Cavium ARM-based processors to its public cloud platform by the end of next month.

The company, an early adopter of the Power architecture for cloud, said it will add Cavium’s flagship 48 core 64-bit ARMv8-A ThunderX workload-optimized processor to its RunAbove public cloud service.

Google Cloud adds Microsoft support as Windows Server 2003 reaches EOL

Google made Windows Server support generally available this week

Google made Windows Server support generally available this week

Making good on commitments the cloud provider made in December last year Google has announced general availability of Windows Server on the Google Cloud Platform. The move comes the same week Windows Server 2003 reached its end of life.

“Making sure Google Cloud Platform is the best place to run your workloads is our top priority, so we’re happy that today Windows Server on Google Compute Engine graduates to General Availability, joining the growing list OSes we support. We’re also introducing several enhancements for Windows Server users,” the company said in a statement on its cloud blog.

“With its graduation to General Availability, Windows Server instances are now covered by the Compute Engine SLA. Windows Server users can now easily deploy a server running Active Directory or ASP.NET using the Cloud Launcher, and can securely extend their existing infrastructure into Google Cloud Platform using VPN.”

Google also said customers the purchase GCP support packages can get architectural and operational support for their Windows Server deployments on its cloud platform. And with Microsoft ceasing support for Windows Server 2003 Google is looking to lure in Microsoft developers by committing to support migration to more current Microsoft Server releases (2008, 2012).

In December last year the company announced it would begin offering Microsoft license mobility for the Google Cloud Platform, enabling existing Microsoft server application users to bring their own licenses and apps – SQL Server, SharePoint, Exchange – from on-premise to the cloud, without incurring any additional fees.

As before the move to expand support for the Microsoft ecosystem is likely to come as welcome news to the .NET crowd, which is fairly sizeable. Microsoft commands a 32.8 per cent share of all public web server infrastructure according to W3Techs.

IBM, Mubadala joint venture to bring Watson cloud to MENA

IBM is bringing Watson to the Middle East

IBM is bringing Watson to the Middle East

IBM is teaming up with Abu Dhabi-based investment firm Mubadala Development Company to create a joint venture based in Abu Dhabi that will deliver IBM’s cloud-based Watson service to customers in the Middle East and Northern Afirca (MENA) region.

The companies will set up the joint venture through Mubadala’s subsidiary, Injazat, which will be the sole provider of the Watson platform in the region.

The companies said the move will help create an ecosystem of MENA-based partners, software vendors and startups developing new solutions based on the cognitive compute platform.

“Bringing IBM Watson to the region represents the latest major milestone in the global adoption of cognitive computing,” said Mounir Barakat, executive director of ICT at Aerospace & Engineering Services, Mubadala.

“It also signals Mubadala’s commitment to bringing new technologies and spurring economic growth in the Middle East, another step towards developing the UAE as a hub for the region’s ICT sector,” Barakat said.

Mike Rhodin, senior vice president of IBM Watson said Mubadala’s knowledge of the local corporate ecosystem will help the company expand its cognitive compute cloud service in the region.

IBM has enjoyed some Watson wins in financial services, healthcare and the utilities sectors, but the company has been fairly quiet on how much the division rakes in; over the past year the company made strides to expand the platform in the US, Africa and Japan, and recently made a number of strategic acquisitions in software automation in order to boost Watson’s appeal in customer engagement and health services.

EMEA cloud infrastructure spending swells 16% in Q1 2015

Spending on cloud as a proportion of overall IT expenditure is growing at healthy rates

Spending on cloud as a proportion of overall IT expenditure is growing at healthy rates

Cloud-related IT infrastructure spending in the EMEA region grew 16 per cent year on year to reach $1.01bn in the first quarter of this year, representing just under 20 per cent of the overall IT infrastructure spend according to analyst house IDC.

Spending on IT infrastructure (servers, disk storage, and Ethernet switches) for public cloud accounts for about 8 per cent and private cloud 11 per cent of the overall spend; the firm previously estimated that growth in spending on public cloud would outpace private cloud spending by nearly 10 percentage points (25 and 16 per cent, respectively).

Michal Vesely, research analyst, european infrastructure at IDC said much of the expenditure in Western Europe was fuelled mainly by public cloud and large-scale datacentre installations.

“Private cloud expenditure, especially on premises, on the other hand, is more directly connected to regular IT investments by enterprises,” he explained. “Private cloud spending saw a slower pace as users assess their storage, as well as integrated and hyperconverged systems, strategies. Once decisions are made, we expect another major push in the forthcoming period.”

The firm also said unstable macroeconomic conditions in Southern and Western Europe haven’t adversely impacted spending trends , although on-premise deployments seem to be growing at a slower rate – in part due to an increased shift to cloud. According to the analyst house this shift is in full swing. In April the firm forecast that cloud will make up nearly half of all IT infrastructure spending in four years.

OVH adds ARM to public cloud

OVH has launched an ARM-based public cloud service just 8 months after going to market with a Power8-based cloud platform

OVH has launched an ARM-based public cloud service just 8 months after going to market with a Power8-based cloud platform

French cloud and hosting provider OVH said this week it will add Cavium ARM-based processors to its public cloud platform by the end of next month. The move comes just 8 months after the company added the Power8 architecture to its cloud arsenal.

The company said it will add Cavium’s flagship 48 core 64-bit ARMv8-A ThunderX workload-optimized processor to its RunAbove public cloud service cloud service.

“This deployment is an example of OVH.Com’s leadership in delivering latest industry leading technologies to our customers,” said Miroslaw Klaba, vice president of research & development at OVH.

“With RunAbove ThunderX based instances, we can offer our users breakthrough performance at the lowest cost while optimizing the infrastructure for targeted compute and storage workloads delivering best in class TCO and user experience.”

OVH, which serves 700,000 customers from 17 datacentres globally, said it wanted to offer a more diversified technology stack and cater to growing demand for cloud-based high performance compute workloads, and drop the cost per VM.

“Cloud service operators are looking to gain the benefits and flexibility of end to end virtualization while managing dynamically changing workloads and massive data requirements,” said Rishi Chugh, director marketing at Cavium. “ ThunderX based RunAbove instances provide exceptional processing performance and flexibility by integrating a tremendous amount of  IO along with targeted workload accelerators for compute, security, networking and storage at the lowest cost per VM for RunAbove – into a power, space and cost-optimized form factor.”

OVH is among just a handful of cloud service providers offering a variety of cloud compute platforms beyond x86. Late last year the company launched a cloud service based on IBM’s Power8 processor architecture, an open source architecture tailored specifically for big data applications, and OpenStack.

But while cloud compute is becoming more heterogeneous there are still far fewer workloads being created natively for ARM and Power8, which are both quite young, than x86, so it will likely take some time for asset utilisation (and the TCO) rates to catch up with where x86 servers are today.

AWS and Chef cook up DevOps deal

Chef is moving onto the AWS Marketplace

Chef is moving onto the AWS Marketplace

IT automation specialist Chef and AWS announced a deal this week that would see Chef’s flagship offering offered via the AWS Marketplace, a move the companies said would help drive DevOps cloud uptake.

Tools like Chef and Puppet Labs, which use an intermediary service to help automate a company’s infrastructure, have grown increasingly popular with DevOps personnel in recent years – particularly given not just the growth but heterogeneity of cloud today. And with DevOps continuing to grow – by 2016 nearly a quarter of the largest enterprises globally will have adopted a DevOps strategy according to Gartner – it’s clear both AWS and Chef see a huge opportunity to onboard more users to the former’s cloud service.

As one might expect, the companies touted the ability to use Chef to migrate workloads off premise and into the AWS without losing all of the code developed to automate lower level services.

Though Chef and Puppet Labs can both be deployed on and automate AWS cloud resources the Chef / AWS deal will see it gain one-click deployment and a more prominent placement in its catalogue of available services.

“Chef is one of the leading offerings for DevOps workflows, which engineers and developers depend on to accelerate their businesses,” said Dave McCann, vice president, AWS Marketplace. “Our customers want easy-to-use software like Chef that is available for immediate purchase and deployment in AWS Marketplace. This new partnership demonstrates our focus on offering low-friction DevOps tools to power customers’ businesses.”

Ken Cheney, vice president of business development at Chef said: “AWS’s market leadership in cloud computing, coupled with our expertise in IT automation and DevOps practices, brings a new level of capabilities to our customers. Together, we’re delivering a single source for automation, cloud, and DevOps, so businesses everywhere can spend minimal calories on managing infrastructure and maximise their ability to develop the software driving today’s economy.”

Dev-focused DigitalOcean raises $83m from Access Industries, Andreessen Horowitz

DigitalOcean raised $83m this week, which it will use to add features to its IaaS platform

DigitalOcean raised $83m this week, which it will use to add features to its IaaS platform

DigitalOcean this week announced it has raised $83m in a series B funding round the cloud provider said would help it ramp up global expansion and portfolio development.

The round was led by Access Industries with participation from seasoned tech investment firm Andreessen Horowitz.

DigitalOcean offers infrastructure as a service in a variety of Linux flavours and and aims its services primarily at developers, though the company said the latest round of funding, which brings the total amount it has secured since its founding in 2012 to $173m, will be used to aggressively expand its feature set.

“We are laser­-focused on empowering the developer community,” said Mitch Wainer, co-founder and chief marketing officer at DigitalOcean. “This capital infusion enables us to expand our world­-class engineering team so we can continue to offer the best infrastructure experience in the industry.”

Although the company is fairly young, and with just ten datacentres globally it claims to serve roughly 500,000 (individual) developers deploying cloud services on its IaaS platform, a respectable size by any measure. It also recently added another European datacentre in Frankfurt back in April, the company’s third on the continent.

But with bare bones IaaS competition getting more intense it will be interesting to see how DigitalOcean evolves; given its emphasis on developers it is possible the company’s platform could evolve into something more PaaS-like.

“We began with a vision to simplify infrastructure that will change how millions of developers build, deploy and scale web applications,” said Ben Uretsky, chief exec and co-­founder of DigitalOcean. “Our investors share our vision, and they’ll be essential partners in our continued growth.”

SoftLayer ups RAM, drops storage and compute costs

SoftLayer rejigged its cloud pricing

SoftLayer rejigged its cloud pricing

SoftLayer announced new pricing model it said would make the company more competitive among other cloud providers, in part by not charging for many of the networking costs.

“While other cloud providers advertise “low” prices for incomplete solutions, they neglect to mention extra charges for essential resources like network bandwidth, primary system storage, and support. At SoftLayer, our servers already include these necessary resources at no additional charge,” the company explained on its blog.

“Our new pricing model includes a redeveloped ordering and provisioning system that offers even more granular pricing for every SoftLayer bare metal and virtual server, from the processor to the RAM, storage, networking, security, and more.”

It also announced location-based pricing, meaning the company will uniquely price cloud services based on datacentre location.

Under the new cost model compute (dual Xeon ES-2620 4U processors) and storage costs dropped while RAM prices increased slightly – though the company said users can expect to save close to 40 per cent overall.

The latest round of cloud cost cutting follows similar moves from others to strip out fees and drop cloud service prices. AWS, Google and VMware have all adjusted their pricing downward in the past few months.