Read this

Winners and losers in the new cloud revolution

Cloud computing is becoming as important as water or electricity. But which companies stand to gain from this shift, and which will lose out? Read ZDNet's complete guide.

• Read more

The range of PCIe-flash drives, named ioScale, can only be bought in batches of 100 or more and capacities vary, per drive, from 410GB up to 3.2TB.

The drives also come with Fusion-io's lowest price, per gigabyte, of flash to date. If customers buy 100 of the largest 3.2TB ioScale drives, they get 3.2PB of flash storage at a cost per gigabyte of $3.89 (lower capacities increase the cost per gigabyte). The price includes all software licences along with the hardware, and comes in at $1.24m for the 100 units, or around $12,000 per drive. This is the official list price though, Fusion-io's CEO David Flynn, says, and can drop further after customers hammer out deals with the company.

So, who has the cash and low-latency storage requirements to shell out over a million dollars for 50TB to 3.2PB of server-attached multi-level cell flash storage? The answer is companies like two of Fusion-io's top customers, Facebook and Apple, Flynn says.

"These IO-memory-based devices are designed for the largest hyperscale customers," he says. "In the hyperscale market customers want to buy more wholesale pricing; we have to date not introduced products specialised for that market."

Fusion-io has been moved to introduce volume pricing for its own products due to the buying habits of such customers. Hyperscale (that is, absolutely huge websites with a need for low-latency server attached memory) tech properties "don't buy through the enterprise server vendors", Flynn says.

It's a trend that's been confirmed by tech execs themselves,  with all the major cloud companies prefering to go direct to the manufacturing source - typically an Asian device manufacturer like Wistron or Foxconn or Quanta.

In addition, recent figures from analyst house Gartner have shown that traditional OEMs have struggled to increase the amount of servers they ship each quarter, while their Asian rivals have increased marketshare.

With all these factors in mind, Fusion-io's move makes sense: if the destiny of hardware is to be bought in commodity volumes at commodity prices, then you might as well try and get a piece of the action. 

"The efficiency of the supply chain is an important part of the ioScale story" — David Flynn, Fusion-io

To that end, the ioScale devices are essentially re-engineered versions of Fusion-io's low-cost ioFX products, which use cheaper NAND flash than Fusion-io's other products. 

"One of the important things about ioScale is that we have moved to a model that allows the contract manufacturers to do things in a more turnkey fashion. The efficiency of the supply chain is an important part of the ioScale story," he says.

The cost-per-gigabyte of $3.89 for 100 3.2TB ioScale units compares with a cost of around $11 per GB for ioDrive2 and under $5 for single ioFX units (Fusion-io would not to disclose the exact price.)

Each ioScale card has a minimum read bandwidth of 1.4GBps and a minimum write bandwidth of 700MBps. The write rate climbs with capacity so a 3.2TB card can write at 1.3GBps.

Read this

Cloud computing's utility future gets closer

• Read more

The virtualisation company instructed its lawyers to take control of open-source project Vert.x from its lead developer and former VMware employee Tim Fox, after Fox left VMware for its rival Red Hat. 

"On the 28th December I received a letter from VMware lawyers (delivered to my door in person, no less!) that I must immediately give up and transfer to VMware all administrative rights over the following things: The Vert.x github project, the Vert.x google group, the domain vertx.io and the Vert.x blog," Tim Fox wrote in a post to the Vert.x Google Group on Tuesday.

To avoid litigation, Fox has duly transferred ownership of the Vert.x domain, blog, Google Group and github organisation page to VMware. 

Vert.x is an application framework that runs on top of JVM (Java Virtual Machine). It lets developers use Ruby, Java, Groovy, JavaScript and Python to write multi-language applications. 

"For now, I will continue leading the Vert.x community the best I can under these restrictions, but we, as a community, need to consider what this means for the future of Vert.x and what is the best way to take the project forward."

Fox had administered the Vert.x project while at VMware and planned to continue doing so at Red Hat. 

After Fox informed the community, numerous members spoke out — mostly wishing to know whether someone from VMware would be taking over administrative duties. 

Joint statement

VMware and Red Hat responded to the community concern with a joint statement posted in the Google Group on Wednesday afternoon, that sought to assuage concerns and stressed that both companies think Fox's continued status as project lead is "an essential component to the success of the project.

"It's important that we try to allay any fears and uncertainty that the community has about the vert.x project and state clearly that VMware and Red Hat are still very much in active discussion regarding how best to support the vert.x project going forwards," Mark Little, a vice-president at Red Hat, and Alexis Richardson, a senior director at VMware, wrote in a joint post to the forum. 

"VMware and Red hat are still very much in active discussion regarding how best to support the vert.x project going forwards" — joint statement

The companies are seeking input from the community about the future direction of the project and whether to move it to an open-source software foundation. 

"Many thanks to all of the input — it has not gone unnoticed," the companies wrote in the post. 

A similar framework to Vert.x is Node.js, whose development is led by employees of cloud infrastructure provider Joyent. 

When companies joust for control of open-source projects, the developer community tends to get nervous, as happened with Oracle's takeover of Java via its acquisition of Sun Microsystems in 2009. 

Although there are continuing concerns about a large proprietary-minded company running a massive open-source project, Oracle has so far stayed true to its goal of being a steward for the Java community rather than a dictator. 

Judging from Red Hat and VMware's post in the forum, the two companies are aware of this worry and are working hard to allay concerns and keep the project free of interference.

Neither VMware or Red Hat responded to ZDNet's requests for further information. 

The company has plans to out-source the manufacturer of some of its storage products to Taiwan-based original device manufacturers like Quanta Computer and Wistron, Digitimes reported on 2 January (subscription required).

Read this

Inside Facebook's lab: A mission to make hardware open source

• Read more

Facebook told ZDNet in September that it uses Wistron subsidiary Wiwynn along with Quanta for the production of some of its stripped-down Open Compute servers, and hinted that storage would be next to be outsourced.

Put this together with the Digitimes report and it is possible that Quanta and Wistron could be bidding to manufacture Facebook's low-cost 'Open Vault' storage array.

By sidestepping traditional enterprise IT companies like HP, Dell and IBM, any sufficiently large company can get IT gear cheaper if they go straight to the production source. All of Google's servers and 30 percent of Amazon's are made by Asian companies, Digitimes reported.

This shift is already having serious repercussions for OEMs, with recent Gartner figures showing that over the past three years traditional server companies have had trouble preserving their market share by units shipped, while low-cost Asian manufacturers have dramatically grown theirs. 

But as with any commoditised business, much of the money to be made in servers and storage is in software, rather than hardware. For this reason major OEMs like IBM still take the lion's share of revenues by using their hardware as a Trojan horse for huge families of lucrative software.

On Thursday morning multiple Azure users began reporting problems with the company's Ajax Content Delivery Network, a technology that helps to quickly serve third-party javascript libraries, such as jquery, to sites and applications built on top of Windows Azure.

As of 1:14pm GMT the ajax.aspnetcdn.com Microsoft site was completely unavailable.

Cedexis, a company that operates a variety of cloud monitoring technologies, also confirmed the outage via its Cedexis Radar technology in a post to Twitter at 2:12pm, which showed a loss in capability of the CDN (pictured).

"Dammit! ajax.aspnetcdn.com is down, @Microsoft any ideas when it will come back up? #downtime#ajax#jquery#CDN," Twitter user Chuck Neely wrote in a post to the microblogging network at 12:49pm GMT.

At the time of writing Microsoft had not responded to a ZDNet request for comment. A trawl of the technology giant's public social media accounts showed no acknowledgement of the issue. 

Read this

Cloud computing's utility future gets closer

• Read more

On Thursday Heroku updated its platform-as-a-service (PaaS) cloud to support version 3.3.0 of the Python runtime as well as 2.7.3. 

Python is a popular tool in the modern web developer's arsenal; in January the PyPL Popularity of Programming Language Index named it the second most popular language in the US behind Java, according to Google Trends data. 

The latest version of the Python runtime differs dramatically from Python 2.7.3 with many commands altered to extend Python's cherished simplicity and brevity. 

But only some developers will be able to get the most out of the latest runtime as many Python libraries are yet to be updated from the second generation, so some developers could struggle to get their code up and running without having to recode libraries to version 3.x

Besides Heroku, Microsoft's Azure and Google's App Engine are capable of running Python applications, though both of these platforms only support version 2.7.3 of the language.

Heroku's biggest rival, Oracle-backed EngineYard, does not support Python.

Python 3.3 was released in September 2012, and 2.7.3 came out back in April. 

The acquisition was announced by Oracle on Thursday and sees the company continue its strategy of building cloud-based software packages targeted at specific industries.

Read this

Winners and losers in the new cloud revolution

Cloud computing is becoming as important as water or electricity. But which companies stand to gain from this shift, and which will lose out? Read ZDNet's complete guide.

• Read more

"Modern marketing practices are driving revenue growth and it is a critical area of investment for companies today," Thomas Kurian, Oracle's executive vice president of Development, said. "Eloqua's leading marketing automation cloud will become the centrepiece of the Oracle Marketing Cloud."

Eloqua's board has unanimously approved the transaction. The company's senior vice president of professional services, Karen Pisha, spent nine years at PeopleSoft, which was acquired by Oracle, while chief marketing officer Heidi Melin had previously worked at Taleo, another Oracle buy. 

The transaction is expected to close in the first half of 2013 and specific financial terms were not disclosed. Oracle did not say how long it will take to integrate Eloqua's technology into its Oracle Marketing Cloud product.

The cloud buy fits in with Kurian's vision, outlined at this year's Oracle OpenWorld, of the company offering its technologies "to any customer, any user, any partner anywhere in the world through the internet browser".

It also serves as a reminder of Oracle's fondness for acquiring companies and rebranding the technologies under the Oracle brand name, which plays into the argument that Oracle is attempting to become 'the Apple of the enterprise'.

Dell announced the updates to the Active Infrastructure gear at its Dell World conference in Austin, Texas on Thursday.

Read this

10 ways cloud computing will change in 2013

• Read more

The upgrades see the IT giant integrate technology from its acquisition of Gale Technologies into the appliances' Active System Manager software, along with adding 'tier zero' (memory-based) storage to its servers for fast data access.

Active System Manager can now deploy virtual clusters of compute and storage in an infrastructure-as-a-service manner thanks to the Gale Technologies integration, cutting the time it takes to deploy virtualised servers from days to hours, according to Dell.

The 'tier zero' storage, meanwhile, lets administrators put cached data onto a very fast memory layer that plugs into the brain of the server (the processor) via the PCIe bus, making applications far more responsive.

This technology sees Dell use its own software along with hardware from an undisclosed vendor (though ZDNet understand flash cards are from LSI) to offer a fast storage layer, reducing its dependence on technologies from other vendors like Fusion-io and EMC.

Taken together the updates reflect Dell's growing enterprise ambitions and build on the first generation of Active Infrastructures, which were released in October.

The systems will see Dell compete against integrated appliances from HP, IBM, Cisco and - to a lesser extent - NetApp.

Read this

10 ways cloud computing will change in 2013

• Read more

The company said its upcoming public and private cloud products will be built around OpenStack, a package of software for running clouds that has received broad backing from the technology industry from companies such as HP, Cisco, IBM, Red Hat and Intel.

"Dell is increasing its commitment to OpenStack as its open-source cloud platform of choice for public and private cloud," Dell said in a statement on Wednesday.

Dell said it will also offer other cloud products based on other technologies, such as VMware.

As of Wednesday, customers keen to experiment with Dell's own flavour of OpenStack for private clouds can email the company to request a preview of 'Dell Cloud Dedicated', the company said.

Dell's decision to put OpenStack at the heart of its cloud technology follows similar moves by HP and Rackspace.

The OpenStack initiative was launched in mid-2010 with code contributions from NASA and Rackspace. Dell, along with other large companies such as AMD and Citrix, has been involved in the scheme from the start.

However, besides Rackspace there are few major clouds today running on OpenStack. HP's cloud is still in a limited beta and there have been few reports from users about how well OpenStack holds up under the pressures of scale and complexity that any cloud entails.

The HP FlexNetwork Utility Advantage Program was announced on Wednesday and sees the IT giant try to tempt service providers into using its gear and software via a pay-per-use financing programme.

Read this

Cool runnings: IBM's recipe for a happy datacentre, in pictures

• Read more

Switzerland's incumbent telco Swisscom has already signed up to the programme.

"Our customers want to focus on running their business rather than operating networks for employee access to applications like voice and videoconferencing," Oliver Spring, head of product line management for Swisscom, said in an HP statement.

"With HP's FlexNetwork Utility Advantage Program, we can close the gap in our managed service offerings to deliver a networking infrastructure that allows our customers to move beyond operations."

The programme sees HP equip service providers with HP software and hardware based around the company's HP FlexNetwork architecture, which they can then use to offer managed services to customers.

The customer only pays the service provider for what they use. The service provider rents the hardware and software from HP on a month-to-month basis, and so they pay no upfront costs. 

The move suggests that HP believes that service providers are going to look at providing their own cloud services to businesses in a bid to maintain competitiveness against Amazon.

IBM launched a similar financing scheme in September in which it offered service providers financing and support for using IBM gear.

One way for a tech titan to deal with a pesky start-up is to spread FUD (Fear, Uncertainty and Doubt) around the challenger's technology. Another approach would be for the incumbent to invest some of its ample resources in dominating fringe markets into which the smaller company would like to expand, shutting off its sources of future revenue.

Intel (2011 revenues: $54bn), gave details on Tuesday of its latest low-power 32nm Atom processors, the 'Centerton' S1200 series. These chips will go up against ARM (2011 revenues: $785m) in dense, low-power 'microservers' in modern datacentres.

Closing power gap with ARM

The 64-bit S1200 series range in power consumption from a thrifty 6.1W up to 8.5W, have clock rates of 1.6GHz to 2GHz, and have PCIe 2.0, along with support for enterprise features like ECC and virtualised workloads.

On paper, these chips come within a hair's breadth of current ARM Cortex A9 datacentre chips, such as those used by Calxeda in its servers. Calxeda's server nodes have a TDP (thermal design point) of 5W.

Calxeda's figures incorporate the power footprint of memory and the networking interface card (NIC), along with the processor, so the real power consumption of the processor itself is a tad lower. Intel, meanwhile, is reporting TDP for just its CPUs — no NIC and RAM included.

This illustrates how keen the company is to put out numbers that indicate its chips have a headline TDP figure that's competitive with ARM.

Intel was not able to provide directly comparable figures showing the TDP of the S1200s with networking and memory factored in.

First low-power 64-bit chips to market

However, Intel will be first to market with 64-bit-capable chips, as ARM's 64-bit designs are not expected to become available until mid-2013. Applied Micro should deliver the first 64-bit ARM chip in mid-2013, while Calxeda expects to get 64-bit ARM servers out in early 2014.

Plenty of Intel OEMs announced plans for Centerton servers on Tuesday. A full list was not available at the time of writing, but promotional Intel materials seen by ZDNet included companies such as Quanta, HP, Huawei, Dell, Wiwynn and Supermicro among the 20 'design wins' the new Atom chips have racked up.

In 2011 Intel predicted microservers could make up to 10 percent of the server market, and the company is sticking by this number.

The leading proponent of microserver technology to date has been SeaMicro, a start-up server maker that received heavy promotional support from Intel until it was bought by x86 rival AMD. Other companies have dipped their toe in the microserver waters, including HP.

But what does this mean for MY datacentre?

So, if you are contemplating a microserver based around many low-power processors, should you go with ARM or Intel?

Even though Intel's chips consume a bit more power than ARM's, they have the advantage of being based on the well-supported x86 architecture

Read this

Inside ARM: The British success story taking the chip world by storm

• Read more

This means that enterprise applications will work as expected and no code porting will be required.

On the other hand, if companies have built much of their software infrastructure around the LAMP (Linux, Apache HTTP Server, MySQL, PHP) stack, then little code porting will be required for their applications, which could run well on ARM-based microservers.

One benefit of using ARM chips is that if you have a massively distributed software system, such as one using the Hadoop file system and MapReduce information cruncher, then the nature of the platform means that it could cost less to run it on many low-power ARM chips than on a few power-hungry Xeon processors.

However, Intel has been in the datacentre business for decades, whereas ARM licensees are only just beginning to make a mark. Therefore it's likely that many enterprises will opt for Intel microservers first due to software support and vendor confidence.

Next year Intel will bring out another Atom chip — codename 'Avoton' — on its 22nm 'tri-gate' chip fabrication process, which should help cut power consumption further, and in 2014 it hopes to make a 14nm chip.

ARM licensees, meanwhile, will probably use existing 32nm processes from chip fabrication specialists GlobalFoundries and TSMC for next year until those companies' 22nm methods mature. For this reason it's likely that, although it may have a power edge now, ARM's advantage could fade in the next few years.

Read this

Cloud computing's utility future gets closer

• Read more

But accessing these APIs and working out their capabilities can be difficult - an issue that application integration company Mulesoft is hoping to tackle with the launch of the APIHub on Tuesday.

"Discovering, learning, testing and using APIs is still a challenging concern for developers," says Ross Mason, the founder and chief technology officer of the San Francisco-based company.

APIHub will provide a centralised view of APIs, according to Mason. "Comparing and rating APIs is very difficult to do right now, it's very opaque."

The service aims to be to APIs as the Yellow Pages is to phone numbers - with a few added capabilities.

For consumers and developers, it provides a catalogue of APIs along with an interactive development environment for running basic queries. For API providers, it offers a publishing platform and various tools to automatically generate create API documentation.

APIHub aims to have information on 80 percent of modern APIs, Mason says, by using Mulesoft's AnyPoint technology, which sits at the heart of its software tools to help automatically link disparate applications together and has helped the company gather information on thousands of APIs. 

"Discovering, learning, testing and using APIs is still a challenging concern for developers" — Ross Mason, Mulesoft

The service will be available in two flavours - a free website for use by the public, the open-source community and more open companies, and a private version that could be used by a company's internal developers for giving information on proprietary internal APIs - a good use case would be a financial institution.

As for the competition, a company named Mashape produces a similar service via its "cloud API hub" website. 

However Mulesoft's website will be free for both developers and API producers, whereas Mashape puts emphasis on providers using the site to monetise their API by adding billing options, such as charging a certain monthly fee for 10,000 monthly queries.

"The business model is very different," Mason says. "In terms of their platform there is a technical difference [and] if you look at their repository, it's a lot of long-tail small apple APIs."

Mulesoft hopes APIHub will, if it becomes popular, expose developers and companies to Mulesoft's technology and increase the likelihood it can sell application integration software to them. 

Read this

10 ways cloud computing will change in 2013

• Read more

Many cloud computing vendors offer big-data tools that, paired with the ability to rent scalable on-demand compute and storage resources, can provide a potent on-paper justification for analysing data in the cloud.

However, the costs and time it takes to upload large quantities of data into and out of a cloud could mean that businesses that crunch datasets could be hit by unforeseen costs.

"Data has a lot of inertia," Charles Zedlewski, Hadoop-vendor Cloudera's vice president of products, says. "If you run [Hadoop] in the public cloud, that application is going to spit off data. If your data is generated in the [cloud] datacentre, your Hadoop [cluster] is going to be in that datacentre."

Hadoop is a popular open-source platform for data storage and big data analysis. It is based on technologies originally developed by Google to help the search giant store and query massive amounts of data.

Cloudera aims to be to Hadoop as Canonical is to Ubuntu - a proposition that saw the company close a $65m funding round on Thursday.

Location, location, location

Hadoop is representative of a problem facing the big data industry as a whole: where you locate your data storage and analysis engine will determine where the data ultimately resides. And it's for that reason that choosing to locate your big data in the cloud, or in your own datacentre, could have huge ramifications down the line.

For example, if a company wants to analyse its customer data, it could buy in several low-cost servers with large amounts of storage based on chassis designs from companies like Supermicro and run a Hadoop cluster on top. That would give the company control over its infrastructure, where its data resides and the cost of its kit.

However, if the business gets a sudden spike in data that it doesn't have the capacity to process in a timely manner, it will need to kick this data up into the cloud to process and analyse it as an entire set.

For that, it will pay the typical charges to your ISP, along with the fees for renting the associated storage and compute in Amazon, Google, Microsoft or other vendors' public clouds. Upon completing the processing, it may even have to pay additional charges to get data out of the cloud and back into its datacentre.

Data gravity

This is an example of an effect known as "data gravity", which has been outlined by researcher and former EMC employee Dave McCrory. Put simply, data gravity means that the infrastructure where you perform actions upon a dataset will attract more and more data over time and get more and more difficult to drastically change.

"Where public cloud shines is the more moderate amounts of data, where you care about bursting and care about expandability" - Charles Zedlewski, Cloudera

For some web applications it will make sense for data analysis to be done in the cloud, but for others the value is doubtful. This example illustrates why data location can have a big impact on companies' bottom line.

"Where public cloud shines is the more moderate amounts of data, where you care about bursting and care about expandability," Zedlewski says.

Workloads that should be kept in the datacentre are those that generate a huge amount of information that will need to be repeatedly worked on and enlarged, such as those that convert information from the physical world into digital information, like genetic sequencing.

For example, a small sequencing machine can generate a minimum of a terabyte of information for each operation. This data then needs to be fed through the network and into Hadoop, where it is stored then worked upon. Companies that do this kind of work have many of these machines running in parallel. Uploading, processing and downloading terabytes of information from a remote public cloud is not a trivial matter, and the costs and time expenses can be great. 

"The rental fees you're paying in public cloud are five-10 times higher" than on-premise, Zedlewski says. (Cloudera's distribution of Hadoop is run both in public clouds and on premise.)

It all goes to show that although the economics of rentable technology are clear in some cases, such as for scaling websites or one-time number crunching, in others they can be rather cloudy.

Read this

10 ways cloud computing will change in 2013

• Read more

According to the digital cognoscenti, this breed of cloud computing is the way of the future. But what if there was another way to create a global cloud, one that didn't involve cementing the dominance of global technology companies like Microsoft, Google and their ilk? Would you use it?

Symform thinks you would. The three-year-old company told ZDNet this week that, as of 30 November, it is storing petabytes of information across 175 million files in a global peer-to-peer cloud. This data is not being stored in datacentres - as it is in the Amazon, Google and Microsoft clouds - but on the drives of servers, desktops and NAS boxes in 160 countries across the world.

The company is far from being the first to try to put the spare compute and storage capacity of people's machines to work, and judging from those who have gone before it, there's no easy path ahead.

P2P's mixed bag

Peer-to-peer architectures for computing have had variable success over the last decade.

There have been successes, like the distributed computing programs of SETI or Folding@home which use the spare compute cycles on people's machines to search for aliens or sequence proteins, but there have also been failures.

One would be Wuala, a P2P file storage technology developed by researchers at the University of Zurich that was spun off into its own company and subsequently brought by consumer storage giant Lacie. However, shortly after being bought, the peer-to-peer technology was shut down and the system was re-designed to operate from Wuala's datacentres.

Another service based on similar technology was All My Data, but that ran into funding difficulties and shut in 2010.

The majority of Symform's users are IT professionals or IT service provider companies

"The problems in all of these cases is an average consumer is not sophisticated enough and doesn't have enough data," Praerit Garg, the president and co-founder of Symform, says. "The complexity of the problem they tried to solve was too high."

To that end, Symform makes users sign the equivalent of a service-level agreement that sees them agree to provide 80-percent uptime from their Symform hardware over a two-month period. The majority of Symform's users are IT professionals or IT service provider companies. This strategy works because IT professionals are more likely to want a cloud data storage alternative to Amazon Web Services or another major cloud than a consumer, Garg says. 

These companies and individuals are "afraid of cloud because it marginalises their role. Here is a system they can be part of", he says. 

But will P2P be the success that its proponents hope?

"The challenge with P2P storage products is that the cost of storage is falling precipitously, making P2P storage often no more cost effective than cloud storage solutions," the chief executive of a cloud storage company who wished to be anonymous told ZDNet. "I have nothing against the model, it just hasn't worked out well in the past."

Read this

AWS: The complete guide to Amazon's cloud

• Read more

The update was announced on Friday and lets administrators sync data between different datacentres across the world via standard TCP (secured with SSL encryption). The Riak database is fully compatible with Amazon Web Services's S3 storage APIs.

Combined with multi-datacentre replication, the technology helps companies build on top of Amazon Web Services while protecting their data against failures by replicating it across continents, avoiding the problems that befall companies that only locate their data in one of Amazon's global datacentre hubs. 

"Customers of RiakCS fall into one or two characters... enterprises which are mostly internal,
and service providers," Andy Gross, Basho's chief architect, said. "For the enterprise it [gives] redundancy, fault tolerance, disaster recovery [and it] lets service providers offer a true global scale cloud offering."

RiakCS has two data replication options for cloud administrators: full sync and real-time sync. Full sync copies data from a primary RiakCS store to a secondary site at a frequency of administrators' choosing, though the default is six hours. The secondary data stores regularly ask the primary datastore whether anything has changed and, if it has, they will update their own data to bring it in line.

Real-time sync, meanwhile, triggers when a person requests information from a RiakCS pile of data. If they are requesting from a secondary site, the database will check with the primary to see if anything has changed and update accordingly, while if they are requesting data from the primary, there's no wait.

Read this

Losers in the cloud revolution

• Read more

Though the technology has a number of handy features it also highlights the technological gulf between small start-ups and some of the world's largest cloud providers.

Google's Spanner globally consistent database allows for multi-datacentre replication but cuts the need for individual datacentres to talk as much with one another by using a Google-invented API for the passage of time, known as TrueTime, which depends on GPS receivers and atomic clocks.

This lets Google respond to local datacentre requests much faster than other technologies some of the time, as it means in data-read cases it sometimes does not need to sync globally.

"Our solution is probably simpler than Spanner," Gross said, adding: "it's not synchronous, but it is a similar capability."

For now unless you work at Google you can't design applications on Spanner, so Riak may be a good bet. However, given Google's aggressive moves with its Google Compute Engine cloud platform against Amazon Web Services, there is a chance that it could start offering the technology soon.

The latest report by the analyst company on the worldwide server market was released on Wednesday. It paints a grim picture of the global datacentre market, with the market's overall revenues during the third quarter decreasing 2.8 percent year-on-year and overall shipments growing a measly 3.6 percent.

"Server revenue was weak due to ongoing economic weakness and market segment differences," Jeffrey Hewitt, a research vice president at Gartner, said.

If you look over the past three years of data from Gartner, these "market segment differences", appear to come from growing competition from smaller or specialist vendors, with much of the focus on low-cost servers for huge clouds. 

Gartner only lists the top five server vendors by revenue and shipment each quarter and groups the rest into an "other vendor" section. When you look at which companies manage to climb into the fifth position a picture emerges of who these "other vendors" are. Having looked at three years of data we know they include Lenovo, Oracle, Cisco and NEC.

Low-cost vendors

In addition, low-cost Asian device manufacturers like Huawei, Quanta and Wistron are all gaining in importance as well, according to Adrian O'Connell, a research director at Gartner.

Read this

Winners and losers in the new cloud revolution

Cloud computing is becoming as important as water or electricity. But which companies stand to gain from this shift, and which will lose out? Read ZDNet's complete guide.

• Read more

"What we're seeing with this growth in what we categorise as 'other vendors' is some of the Asia-Pacific guys - Lenovo, Huawei - but a big part of it is what we'd refer to as the 'hyperscale class' of datacentres," O'Connell said. 

In other words, as the Other Vendors (pictured above, green) category grows, the more servers are being shipped from low-cost companies to the Facebooks, Googles and Amazons of the world. (Facebook, for example, recently outfitted one of its new datacentres entirely with equipment made by a low-cost Asian manufacturer.) Meanwhile, the typical OEMs have not managed to increase shipments. 

This adds to a pile of ever more substantial evidence that indicates OEMs will lose out from the rise of the cloud. 

Though the server old guard are still making most of the cash - IBM took in almost $3.5bn in the third quarter, compared to $2.6bn for the entirety of the "other vendors" group - they are not shipping nearly as many servers.

Compare 2008 to this year: in the third quarter of 2008 the "other vendors" shipped 629,864 bits of kit, while market-leader HP sent out 724,024. In the third quarter of 2012, by contrast, the "other vendors" managed to ship 846,734 servers between them, compared to 634,793 for HP, which is still the largest single shipper of servers.

Giants struggle

This plays into a trend that has been happening for the last few years - the fringe "other vendors" are steadily shipping more and more kit each year, while the typical enterprise giants of Dell, HP and IBM struggle to meaningfully grow the number of boxes they ship. Along with this, many of these "other vendors" are cheaper than the major OEMs, thereby putting downward price pressure on the entire market. 

OEMs really are looking set to be the losers in the cloud revolution

Gartner's figures also demonstrate how new technologies are changing the server market. Many of the innovations developed by cloud providers and software companies, these days, are based around using software to run jobs across many individual servers, allowing organisations to buy several cheap servers from low-cost vendors instead of one expensive (and powerful) bit of kit from a typical OEM. Though this is most common in large clouds, like Facebook, it also happens at smaller companies. 

As far as the data indicates, it's a trend that's only going to accelerate. Evidence, if it were needed, that OEMs really are looking set to be the losers in the cloud revolution.

Read this

Winners in the cloud revolution

Some companies stand to benefit from the rise of the cloud and others stand to lose. ZDNet picks the IT winners.

• Read more

The update was announced on Monday along with a range of other improvements to the company's cloud services.

"Now, customers using Google App Engine, Google Cloud Storage, Google Cloud SQL and (soon) Google Compute Engine can deploy their applications, data and virtual machines to European datacentres," the company wrote in a blog post. "This helps bring your solutions even closer to your customers for faster performance and enables international redundancy."

Google will not let customers select exactly where in Europe their data will be stored, potentially locking out customers who must keep their data within the legislative bounds of their country. 

Other cloud companies have also tussled with this problem - Amazon Web Services has an Irish datacentre hub and several edge locations across Europe. Microsoft lets users put their data in two European datacentre hubs: Ireland and the Netherlands. However, the company does have the ability to "transfer customer data within a major geographic region for data redundancy or other purposes", Microsoft says.  

Though Europe has lagged behind the US in cloud uptake, Google says that European datacentres from big cloud providers may boost overall European cloud consumption.

"I think that the fact we're offering the extended European datacentres has in a way boosted the adoption of cloud in Europe," Barak Regev, head of Google's Cloud Platform in EMEA, told ZDNet. "Making all of our services available through EU datacentres will accelerate that even further."

Read this

AWS: The complete guide to Amazon's cloud

• Read more

The additions to Google Compute Engine, the company's infrastructure-as-a-service technology, were announced by Google on Monday.

"At launch we offered four basic [instance types]," Barak Regev, head of Google's Cloud Platform in EMEA, told ZDNet. "If you aggregate [the 36 new instances] into brackets, you'll see they are touching the high-memory and high-CPU instances [but] the one that is most notable is the diskless file configuration."

The diskless configuration's instances let developers rent lower-cost instances for applications that do not need a dedicated disk attached to their server - an 'ephemeral' disk - but can handle a separate 'persistent' disk. Ephemeral disks are faster but do not retain data across instances. 

Google is also cutting the prices of its main four instances by around five percent.

Alongside the compute upgrades Google is also preparing to launch a technology named 'durable reduced availability' storage as a limited preview. This product reduces the cost of storing data but with lower availability. It is roughly analogous to Amazon Web Services' existing S3 Reduced Redundancy Storage. Google did not disclose by how much cost would be reduced.

"What we are doing in Google is continuing the momentum of [our] cloud platform" — Barak Regev, Google

Furthermore, Google is adding 'persistent disk snapshotting' so that administrators can create a backup of their persistent stored data and then move it to other Google datacentres around the world to help create multi-continent applications.

Asked when Google expected to have technology parity with Amazon Web Services, Regev noted that Google has some tech - like BigQuery - that Amazon lacks, however he hinted that Google is due to make several announcements that will close the gap in other areas.

"Obviously people are looking at the more popular offerings like Amazon or other companies; what we are doing in Google is continuing the momentum of [our] cloud platform," he says. "I think that today's announcement together with some upcoming announcements will change that perception we might have [as lagging]."

The acquisition was announced by Dell on Friday. Alongside this, the company said it is forming an Enterprise Systems and Solutions group, which will help with its strategy.

Santa Clara, California-based Gale Technologies was founded in 2008 and has two main products: GaleForce, which manages and automates datacentre infrastructure; and Cloud Scheduler, which helps administrators provision, monitor and improve the efficiency with which virtualised equipment works.

These products will bolster Dell's cloud management portfolio, the Round Rock, Texas-based company said.

"Gale Technologies integrates well with our Active Infrastructure family, and provides an intuitive, flexible and comprehensive foundation for application, virtual desktop infrastructure and private cloud deployments for our customers," Marius Haas, head of enterprise at Dell, said in a statement.

The purchase, combined with the new unit, moves Dell further on its journey to become an enterprise services company rather than one known mostly for its hardware.

The seeds for this transition were sown a year ago at the company's inaugural Dell World conference in Austin, Texas, where chief executive Michael Dell outlined its plans.

This strategy bears similarities IBM's shift over the past couple of decades, during which the company has shied away from easy-to-commoditise hardware and moved further into software and services.

However, unlike IBM, Dell remains committed to its personal computer division. The only other major company that has tried to shift into being a services-led organisation while continuing to make low-end hardware is HP, which is beset by wide-ranging problems with a turnaround unlikely to happen for years.

Given this, Dell's new Enterprise Systems and Solutions unit, part of the overall Dell Enterprise Solutions Group, will focus on creating bespoke datacentre infrastructure templates for customers. These will help the company sell a combination of hardware and software into modern datacentres.

Existing utility markets include ones for water, electricity, gas and, to a degree, basic internet connectivity. 

A utility market occurs when an item has been commoditised to the point that it becomes very hard to differentiate on a technology basis, and instead companies distinguish themselves through different levels of service, availability and support.

We already have indicators from the IT industry that the fundamental tools for offering cloud computing are commoditising

In the same way that in the early days of electricity there were arguments over whether AC or DC delivered the 'best type of electricity', the technology industry continues to debate the merits of certain technologies over others for delivering cloud computing. However, these arguments are growing less fervent as datacentre infrastructure is commoditised and homogenised by large cloud providers. 

The consequences of remote storage and compute becoming a utility are very wide-ranging. This market has the potential to build up some companies and crush others, cement incumbents and offer new business opportunities for savvy startups. It will also have a huge effect on how businesses and consumers access and use technology.

We already have indicators from the IT industry that the fundamental tools for offering cloud computing are commoditising. So, what defines a utility market and is the cloud computing getting close?

James Constant, the chair of Energy Forecaster and former Head of Power at Russian utility Gazprom, says the requirements of a utility market are:

• a source of energy generation
• a transportation network
• a transmission and distribution capability
• a metering competency
• a price-setting mechanic
• a regulator to ensure protocol adherence
• a customer 

Along with this, for this market to be competitive, you need:

• a set of willing participants
•  a market mechanism
• a trading methodology

How does cloud compare?

When we set these requirements against the current state of the cloud market, it looks as though cloud is close to becoming a utility.

• Any computer with a capability to transmit information has the ability to be a source of 'energy' - in the cloud, compute and storage - generation
• The internet and its transit providers form the transportation network.
• Datacentres can handle transmission and distribution as they house the compute and storage capacity.
• Public cloud services can precisely meter how much storage or compute is being used at any one time, though different providers can have different pricing methodologies.
• As for price-setting mechanics, the fundamental price of public cloud computing is what it costs to receive a request, process it, and then transmit information back. This is determined by a wide-range of factors, with cloud providers able to directly control the hardware and software cost by engineering their own systems and secondary costs (facilities, staffing, electricity) according to their klout with local suppliers. 
• Score: 5.5 / 7

All that’s lacking for this section is a regulator. Whether the cloud computing industry should be regulated is a complex issue that will undoubtedly become a major debate before long.  

As for the competitive criteria: 

• Google, Microsoft and Amazon all compete for the infrastructure-as-a-service market, while competition is hotting up in the platform-as-a-service area as well. SaaS has a reasonably mature competitive model, though this is less like a utility market due to the variance of the software being delivered. Many industry watchers believe the battle between Google and Amazon will drive the formation of a true utility market due to vicious pricing strategies. 
• There is not yet a clear market mechanism for homogenising compute and storage from different providers making them truly interchangeable. There have been some attempts to create a true market for this - like Spotcloud - though there is little evidence these services have caught on. 
• A trading methodology: We are beginning to see the first signs of this via some companies, such as Strategic Blue, which is a UK-based cloud brokerage that buys cloud computing resource from a variety of providers like Amazon Web Services, Joyent and Firehost, then resells it to IT buyers. The company uses open source tools like Chef and Puppet to abstract the IT environment away from the peculiarities of each cloud to be able to sell a single resource. However, work needs to be done here: “At the moment, [cloud] pricing is not rational,” James Mitchell, the chief executive of Strategic Blue, says. Nonetheless, the existence of a company that can work as an intermediary between clouds and customers implies access to a utility market because otherwise Strategic Blue’s business model would not work. 
• Score: 1.5 / 3

Cloud scores 7 out of 10 overall for qualifying as a true utility. 

How long until utility?

Cloud computing has taken on most of the traits of a utility market, but has not transformed into one yet.

How long might it take for the cloud computing industry to evolve into having a utility service capability?

Based on the current indicators, the two barriers are the lack of a regulator and the inability for cloud companies to trade capacity among themselves. 

Read this

Winners in the cloud revolution

Some companies stand to benefit from the rise of the cloud and others stand to lose. ZDNet picks the IT winners.

• Read more

There is no real technical barrier to Microsoft being able to sell or buy compute and storage from Amazon as needed, or vice versa. However, as all the major cloud providers are fixated on building out their own competing infrastructures to let them maintain price parity with their competition, it is unlikely they will trade capacity amongst themselves directly.

Nonetheless, the existence of forward traders like Strategic Blue shows that there is some kind of mechanism here, though it is an immature one. 

What could force its transition into a true utility could be a datacentre capacity crunch, which would cause cloud providers to trade capacity amongst themselves. However there are few signs that this could happen. 

From a developer's point of view, the cloud already has taken on enough characteristics of a utility that, barring massive problems at any one provider, it behaves like one and looks like one to the IT consumer. 

As cloud computing continues on its path to become a utility, the benefits to IT consumers will grow as prices are successively cut, but companies that cannot operate at the necessary scale of a utility are likely to run into problems. 

Read on to find out how the cloud revolution could make a few large cloud providers tech giants and make life difficult for start-ups. Then discover which companies are likely to win out in the new cloudy world and which may have problems. 

The new storage features include the ability to pool storage into volumes split across multiple virtualised drivers on multiple servers, saving potential spend on dedicated storage gear.

As well as the added storage features, Rackspace's private cloud distribution of Openstack also saw upgrades to its automation and systems management.

The free Rackspace Private Cloud Software, released in August, gained the new features on Thursday following the integration of technologies from the recent Folsom release of OpenStack.

"We did version one [of the Rackspace Private Cloud Software] because we recognised that unless you had an awful lot of Python skills and technical knowledge, getting Openstack up and running wasn't as easy as it could be," Nigel Beighton, international vice president of technoloogy at Rackspace Hosting, said. 

Besides the help-the-developers reasoning, Beighton said Rackspace has also been building out the software so that it can get more people used to working with OpenStack. The theory is that, as more people use the software, there's a chance they will either pay Rackspace for additional support, or move some of their data into Rackspace's public cloud, which is based on the same technology.

While the software is available for free, it is limited to running across 20 servers. To scale beyond that, users will either have to switch to the main open source OpenStack distribution, or if they want to pay for support from Rackspace, pay Rackspace to up the node-limit.

OpenStack is an open source tool for managing and running private and public clouds and is designed to do for the cloud what Linux did for the server - that is, provide a credible, easy-to-use competing set of tools to those made and sold by large companies.

Rackspace co-founded the OpenStack initiative in 2010 along with NASA. Since then over a hundred IT companies have joined and contributed code to the software.