While not at the level of the Tier IV fault tolerant datacenter, the requirements to meet the Tier III concurrently maintainable site infrastructure standard are still substantial; at the very least the datacenter requires redundant capacity components and multiple independent paths serving the computer equipment along with at least 2N power supplies and matching  installation to the site topology.

Digital Realty provides turnkey and colo services at their datacenters and currently focuses on their Turn-Key FlexSM POD-focused modular architecture. By building their facilities out to the Tier III specification level they are able to offer their customers a turn-key facility with high reliability and features that usually require significant advanced planning on the part of the customer.

The first five datacenters to achieve this certification are in Australia (4) and the United States (1). Digital Realty has not announced a schedule for the certification of the 15 additional Tier III certification projects for which they have contracted the Uptime Institute.

Weeks after the new measure was signed into law, Utah Governor Gary Herbert’s staff received an email expressing the NSA concerns about the new tax, pointing out that it came as a surprise and that stable power prices were one of the major factors that led to the selection of the Utah site. The surprise part was itself surprising, as an attorney for Utah stated that the agency had been informed of the proposed tax before the measure had been signed by the Governor.

With an estimated yearly power bill of $40 million, the tax would add an additional $2.4 million to the operating costs of the datacenter. All of the major state level players who were involved in attracting the datacenter project are claiming that they were unaware of the potential tax, nor were any of them consulted about the impact it could have on the datacenter facility. At this point in time, the newspaper is reporting that there are negotiations ongoing about the facility, and its tax liability is scheduled to be resolved by September.

UPDATE: Nate Carlisle, of the Salt Lake Tribune, sent me the following information: The Military Installation Development Authority still has discretion on when, or if, to apply the tax and at what rate up to 6 percent

So at this point the decision if the tax is to be applied, or in what amount, has not been announced.

This is an important moment for AMD; their decade’s long competition with Intel has gone full circle; they compete now based almost solely on price, having seemingly forever given up the technology leadership position to Intel. So with their new designs, focused on building efficient, cost effective, performance capable datacenter servers based on open standards, it could be said that AMD is betting the farm. Given the positive response that there has been, from both vendors and consumers to the Facebook lead Open Computing project, this might be a pretty safe bet.

Designed to fit in any standard 1U to 3U rack height server chassis (both the Open Compute and standard 19” rack models), the motherboard boasts components such as dual Opeteron 6300 Series processors, 12 DIMM slots supporting up to 384 GB of RAM, six SATA ports, integrated dual channel Gb Ethernet, multiple PCIe slots, and a custom connector for modules from vendors Broadcom or Mellanox.

The Open Compute servers are coming from vendors like Tyan and Quanta and through major distributors, but not from traditional first choice datacenter hardware vendors like IBM, HP, or Dell. If anything, the first stirrings of the Open Compute products are in direct competition with the traditional model of datacenter hardware sales and its proprietary nature, despite the much more commodity-like aspects of rack mounted servers.

The new hardware, a cluster of four 24 SOC Viridis high-density ARM servers using Calxeda SOCs built by server vendor Bostion Limited, leveraged the most recent version of Fedora (Fedora 18), which included these standard tools for ARM CPUs. This implementation is a major step toward making ARM a significant part of the core Fedora architecture. It's also the first migration for Project Fedora to a production ARM datacenter-level server, using standard tools and off-the-shelf datacenter hardware.

That's probably the most significant part of this development: The hardware environment that Fedora 18 was moved to is already in commercial production and in use with other versions of the Linux operating system. As Fedora is basically synonymous with Red Hat, this is a direction that was necessary in order for the Red Hat business to remain competitive in all aspects of the future datacenter.

That's pretty much why the goal of this rollout is to accelerate Fedora's community development of the ARM port at the datacenter level, hoping to catch the wave of ARM acceptance as a datacenter server standard. The current hardware is still only 32 bit, however, which means that there will need to be a significant hardware upgrade in the not too distant future as 64-bit ARM SOC hardware begins to appear, allowing for development on that platform, which will likely be the primary datacenter standard.

This determination was reached as a result of an audit of the process, performed over a six-month period covering the end of 2012 and beginning of 2013. With 24 governmental organizations explicitly being targeted for the consolidation and datacenter closure process, the report found that only five of them had reported progress that resulted in estimated datacenter savings. Those groups that have reported closures and estimated savings reported only general information, and the reports lack information on the size of the datacenters being shut down.

The report also highlighted the fact that despite the GAO starting in June 2011, there were no complete plans for the consolidation process submitted by any of the agencies doing the IT downsizing. There had also been little progress made in inventories being completed and plans for the process being submitted. The finding was that in June 2012, only three agencies had submitted complete inventories, and only one had submitted a consolidation plan. And even those three inventories were outdated, as the OMB had changed the reporting requirements, and none of those inventories matched the new model.

Testimony by various responsible parties in front of congress has also focused on a lack of leadership from the OMB and a certain level of concern that the efforts being made aren't targeted at achieving the highest possible level of savings. The GAO has described the level of savings currently achieved as "minimal", and the reports from the affected agencies themselves, which are likely to be optimistic, claim less than $650 million in savings, or just over 20 percent of the goal.

Related stories

• Can government agencies play nice and share?

• Senators attempt to give federal datacenter consolidation the force of law

• New website hopes to improve the Federal datacenter consolidation process

• Federal Data Center Consolidation: Who's paying?

Using a public .NET API and an Operations Manager Watcher node, the MP integrates Amazon's existing CloudWatch information so that it appears as management information, and alerts within Microsoft System Center Operations Manager. The management pack can monitor EC2 instances (Windows and Linux), Auto Scaling groups, Availability Zones, Elastic Beanstalk, Elastic Block Store (EBS) volumes, Elastic Load Balancing, and CloudFormation stacks.

Using this management pack also makes it possible to drill down into applications running on the AS cloud, providing there is an appropriate application management pack for the Application available to MSSC. The MP is also capable of separating the operating system from the virtual machine, allowing problem resolution at a much more granular level. This means that the monitoring console can be used to determine if the issue is in the cloud itself or in the specific OS/application running in the cloud.

If Microsoft is serious about becoming a major provider of services rather than just applications, this type of third-party integration is just the start for Systems Center.

The goal is to build a standardized switch that will allow the consumer to load the operating system and configuration of their own choice, and not be locked into the proprietary OS infrastructure currently found in the network switching market. The purported reason is to allow for the development of networks that will easily scale to meet the needs of large datacenter operators at a price point significantly lower than the per port cot of the cutting edge switching vendors currently in the market.

The idea of replicating the bare metal server concept with the bare metal network switch is an appealing one. Need more capacity or a new project specific configuration? Pull a blank out of stock and drop it in to the rack that needs the expansion capability, then image the entire rack, servers, and switch to address the specific project or need.

On one hand, you would think that network switch vendors would find this prospect an anathema; they are selling specific capabilities and capacities, and charging their customers for the advantages that they claim over their competitors. On the other hand, it would allow a certain level of networking switching to become completely commoditized, and that's where the real issue lies.

There's no real problem with a commodity level network switch. Many, if not the majority of switches are installed to handle the mundane tasks of network operation. Vendors that cover a broad range of switching capabilities already offer relatively inexpensive switches that perform at this level. A standardized level of capability and hardware would simplify this market immensely, both for the vendor and consumer.

But the high-end is where this model breaks down. Custom switching fabrics matched with highly optimized operating systems and management tools are what separate this level of switch from a potential commodity switch. In an environment like Facebook or Google, massive standardization across the datacenter takes a certain amount of priority over optimization. For business like these, the advantages of a standardized/commoditized networking infrastructure are huge.

But for companies who are looking for that competitive business advantage over their market segment, a high-performance, highly optimized networking infrastructure can be worth the investment. Taking advantage of the network switch vendor investment in developing the fastest, most capable switching fabric and matched optimized operating system will be the only way to guarantee the best possible performance from your networking infrastructure.

So while the commodity level open standard switch may well be a successful player in the networking market, the cutting edge, high-end switch will still retain its place at the top of the networking rack.

From the high-level perspective, this is a great idea, and a serious differentiator for ProfitBricks. For people like me, those who keep a close eye on the IaaS market and advise clients on vendor selection, availability should be a top concern. There has been a seemingly continuous series of events, many very high profile, of major services hosted on IaaS platforms going down because of issues that have long been addressed in physical datacenters, but are just seeing processes developed to deal with in the world of the virtualized infrastructure.

ProfitBricks has rolled out its high-availability solution based on Address Resolution Protocol. This allows dynamic routing of IPv4 address to virtual machines. This capability is now standard with its service. ARP is not a new protocol, but ProfitBricks claims to be the first IaaS provider to implement it in this environment, bringing this brick-and-mortar datacenter standard to the cloud. Its cloud will now support industry standard Linux distributions such as Debian and Ubuntu deployment with Linux-HA, without any special configuration requirements.

This may not sound like a big deal, but current HA solutions for cloud vendors have started off by adding levels of complexity that supporting ARP doesn't have. Being able to build HA networks using traditional models and technologies will be important in driving further business adoption of the cloud.

So last year's announcement was a step that many customers felt that Salesforce.com needed to make, but six months later, there didn't seem to be much movement on the new EU location. The word was that the choice of providers had been narrowed down to two candidates, and that a decision was imminent, but no date or provider was confirmed.

Much to the surprise of many, Marc Benioff announced at Dereamforce last year that they had chosen a location and vendor. He also said that the service had been specced out and would be operational next year (2013). Of course, he also announced the caveat that they hadn't actually signed an agreement yet.

Getting that agreement in place clearly took longer than expected, as it was confirmed just today that NTT Europe would be building a new datacenter to collocate Salesforce.com in the UK, and that the datacenter would not be opening until 2014. The planned datacenter, Salesforce's sixth, will be focused on providing its cloud services to Europe, the Middle East, and Africa.

Fuel cells and solar energy have been hallmarks of the most high tech datacenter deployments in recent years, starting with Google adding Bloom Energy equipment as a power source in their California headquarters. Verizon has previously installed fuel-cell power at an office building and switching center located in New York. The solar panels will be installed as rooftop and ground-level sites, and deliver slightly more than 10 percent of the total additional power. Their primary function will be as an add-on power source. Verizon already has significant experience with this process; in various sites in western US states, they have already installed 26 solar power assist systems.

The bulk of the renewable energy will be delivered by ClearEdge Power's PureCell Model 400 systems. These fuel cells will be installed at network and datacenter sites in New York, New Jersey, and California. With over 60 million kilowatt-hours of power delivery, this will be one of — if not the largest — single deployment of fuel cell power systems in the datacenter and networking industry. It's also ClearEdge's largest project delivered to a single-user.

Fuel cells power has gone far beyond simple "proof of concept" projects. As part of Verizon's long-term goal to reduce their corporate carbon footprint, fuel cells are rising to the top of the list in the choices for green, renewable energy.

That's all well and good, but to me, Netflix is not a good example of the direction that cloud services will take, at least from the business side. This is because Netflix is primarily a passive service for the user. The user makes a selection, and Netflix streams that data to the user. Given the size of the data and the amount of internet traffic that Netflix generates, its level of reliability is actually pretty impressive, especially given that they are hosting on Amazon, which means it probably doesn't have the level of control over its back end that it would like (and SJVN's article talks about how Netflix looks to minimize the impact of failures on Amazon's part). But the bottom line is that Netflix is a one-way cloud service, and for the promise of the cloud to be realized, a two-way interactive experience is what will allow users to get the most from the cloud.

And the best example of what can be done with the cloud from this perspective is definitely Valve Corporation's Steam platform. If you're a gamer, you know exactly what I'm talking about. If you aren't, here's the quick rundown.

Steam is a platform for digital distribution and DRM. Users can purchase and download software (not just games anymore). The software can be used in stand-alone modes or linked into the online Steam community, where with more than 50 million users, Steam has seen usage periods where over 6 million active gamers were in concurrent games, interacting with the Steam back end and other users. There is active user support, with close to 150 million user contributions ranging from simple graphic images to full-blown game extensions written with API made available by Valve and other game manufacturers.

Client support includes Windows, OS X, Linux, and mobile operating systems. The overall interface includes software distribution, interactive gaming, software licensing and management, the ability to access product from multiple platforms, an open API for development (Steamworks), social networking, and integration with content that is not Steam specific that the user may already own. The service updates itself and purchased applications, if configured to do so, transparently to the user, assuring them of the most up-to-date experience.

The environment does require downloads to the user's computer, but that doesn't make it any less of a successful cloud-based application service. The size of most games makes download on the fly while playing impractical, a situation that would conceivably be resolved by future improvements in internet bandwidth, but, for most users, that would not be an issue. And in many ways, that's a positive for many people who see that they can still use their applications even if they currently have no internet connection.

Looking at the entire Steam ecosystem gives one a good idea of what the future of cloud computing may bring, with a richly interactive environment that allows millions of users to simultaneously interact with each other, work on their own projects, and be able to transport their environment with them regardless of where they may need to be. And that makes the promise of a cloud future a more realistic and attractive one.

When a company decides to make the choice that colocation is a good addition or replacement for their datacenter operations, they become dependent on their provider's interface to their hardware to manage and monitor their datacenter operations. This means that the provider portal becomes an important part of the day-to-day operations management of the colo customers' IT staff. So the quality of the portal interface and the capabilities for monitoring and management become critical to the success of a colo datacenter customer.

What this has meant to the user is that colo providers have been working hard to allow customers more than just the ability to monitor their equipment. The cutting edge of the business allows deep drill-down into what is going on with the customer equipment and the datacenter infrastructure, along with the ability to control servers, racks, and related equipment in a seamlessly integrated remote solution. Internap is looking to provide a value add for its customers (and a methodology for selling them additional services) by giving the customer the ability to do on-demand provisioning not just of traditional datacenter workloads or equipment, but also cloud service components, allowing for the creation of hybrid colo/cloud configurations.

The cloud provisioning is just another option to the customer that works, from their prospective, the same as the other provisioning capabilities that the portal provides. This process fits in very well with the cloud service model; the customer has the opportunity to expand the capabilities of their business without additional IT infrastructure investment by utilizing back-end cloud services mixed with their existing datacenter back end. The additional services become available to the users of the colo datacenter services in a transparent fashion; they are additional datacenter provided services, and the fact that they are cloud based is neither apparent to the end user nor even necessary for them to know. Management of these services occurs through the existing portal platform management tools, simplifying IT's role in managing the services.

This is one of the simplest ways to make cloud services an integral part of your IT operations. It doesn't relieve IT of the responsibility of selecting appropriate service providers, but is a step toward simplifying the integration of these services, especially for existing customers.

In 2012, AOL was able to decommission 8,253 servers that resulted in savings of nearly $3,000,000. This was after decommissioning 10,000 servers in 2011. Second-place finisher Barclay's, the financial institution, was able to save over $4,000,000 in power and maintenance costs, but decommissioned only 5,515 servers. These were both clearly major restructuring efforts, and the other three finalists, who had submitted their numbers to the competition, were looking at much smaller numbers, with fewer than 600 servers decommissioned and saving ranging from the $100,000 to $800,000 range.

But even the smallest of the finalists, which decommissioned only 387 servers in 2012, still realized saving, on power costs alone, of over $100,000. Considering how tight IT and datacenter budgets are, that is significant cost savings that saves funds to be better applied to improving business performance.

Now, while it is unlikely that you have thousands of servers sitting and idling in your network now, there is a fairly good likelihood that more than 10 percent of your servers are severely underutilized, even in small datacenters. I've recently been shown industry reports that peg the percentage of underutilized servers at an average over 50 percent of the machines in any datacenter. Given the potential cost savings for even a few decommissioned servers, isn't it worth the effort to get a good grip on how the servers in your datacenter are being used?

It seems that a somewhat public claim that the datacenter would give the government the ability to track and store information about every citizen was putting people in a tizzy. Many of the emails I received asked if it was true that the government would be reading their email (this seemed like the No. 1 concern of friends and relatives that like to send me chain emails and old internet hoaxes); a lot simply expressed concern that all this data would be stored and the government would sift through it to discover their deep dark secrets.

Now technically, much of what these people were worried about requires a warrant. Even if it didn't, sorting out the actual criminal information from petabytes worth of trivia is a daunting task, even for today's supercomputers. But what really stood out was how little the average person understands about modern data collection methods and how much data already resides in databases out there that they have freely given up.

I found it really humorous that many of the "oh no, they're reading my email" messages came from people using Gmail. I don't know how many times its been pointed out that those targeted ads that Gmail users get are targeted by using data collection techniques against the contents of the users email.

And a significant portion of the emails I received were from friends and acquaintances that regularly post about their lives, in minute detail, on Facebook. Apparently, they labor under some misapprehension that what they post online remains forever under their control.

And when I asked, just about every person I talked to said "yes" to the question about store loyalty or rewards cards. So after years of giving up detailed information on their buying habits and filling out online surveys, they somehow think that the government is going to find out something new about them that doesn't already exist in a commercial database somewhere.

And these are just the primary ways that most people are already giving up all the details of their life. People write all sorts of disclaimers in things that they put online, as if those "you can't use this information for anything" paragraphs have any real value.

The bottom line is pretty much this: Yes, the government can build facilities that will potentially have the capability to seriously invade your privacy. But why should they bother when the vast majority of the country is already giving up that information freely?

On the European side, the investments are just beginning. At the end of last year, Google announced that it was investing an additional $180 million in its datacenter in Hamina, Finland, just slightly less than its initial investment in the facility with an upgrade project that will roll into 2014.

And yesterday it announced that its $250 million Belgian datacenter, a technical showpiece for energy efficiency due to its novel use of water from the local canal and no traditional chiller system, will be getting an update to the tune of a $390 million investment, more than doubling its initial spending on the datacenter that has become the primary site for European access to the majority of Google's free services.

The Belgian datacenter already employs 120 people, no specific plans have been announced on potential employee increases along with the datacenter investment. Nor has Google announced a timetable for the upgrade process for the facility.

The latest generation of the Wave Glider series of autonomous marine robots is the first to be able to handle more than a pair of computer systems. With expansion capabilities to more than 24 computers and wireless connectivity that can be used to coordinate an entire fleet of these modular computing robots, the possibilities for delivering significant levels of analytical power in otherwise inaccessible conditions and locations has expanded far beyond the capabilities of the previous generation of Liquid Robotics automatons (which themselves set records for time at sea and distance travelled).

The on-board processing power means that these robots are not simply data collection devices, but are also capable of analyzing the data that has been collected and reporting on the results. This is why the company is referring to them as datacenters @sea. CTO Robert Hines, the inventor of the robot, has said that the onboard computing power now possible is equivalent to what was found only in supercomputers from the recent past.

With the connectivity available, these ocean-going platforms, which run the Liquid Robotics Advanced Vehicle Operating System, offer services one would expect in a datacenter, such as multi-tenancy, scalability, and the ability to update software on the fly, without requiring a human in place on the vessel.

Read this

Is it better to own or outsource your data center?

• Read more

The naysayer's generally brought up two arguments: Cost and risk. The first is hard to argue with; if you can't afford to make major changes, then you are unlikely to consider that as an option. But no one told me that they actually took a step back and considered the cost in terms of the overall expenditure over time and the potential cost savings. And look at what gets done for very little extra expenditure; perhaps your next round of updates could be consolidated to a single aisle, rather than scattered throughout the datacenter.

And on that aisle, you could try one of the hot-aisle containment tricks, using inexpensive plastic sheeting and PVC pipes for a frame. Pick a single business process to dedicate that new setup and run it in parallel with the existing system until you're comfortable with its function. Then sit back and compare the efficiencies of that setup to how you were previously running. Getting it working for you may not be as simple as swapping in new blades in servers throughout the datacenter, but it might just get you on the path to a more efficient datacenter overall, with minimal risk and expense.

It also looks like many folks consider this an "either or" exercise, when it clearly isn't. While it might have been framed that way, no one is advocating the wholesale replacement of systems in your existing datacenter; a clean sheet approach is only appropriate when you are creating a new facility (either in an existing space or brand new). But if you aren't taking modular and what is effectively large-scale appliance computing seriously, your business is going to end up way behind the curve.

"We don't do it that way" because it embodies too much risk is the response from someone who hasn't actually looked at all of the issues involved.

The practicality of a major update to your datacenter operations by adding a turnkey system like one of the Microsoft Parallel Data Warehouse appliances or a Hadoop-in-the-box solution can't be dismissed because it represents a major expense or change in the way you do business. If it is the solution that is going to get you the best possible ROI, based on speed in getting the capabilities to your users, improved performance, and better efficiency for the affected applications, it should be considered.

The same applies to modular additions to your datacenter. Perhaps it is time to look at containerized or modular update, allowing you to make significant improvements with having a negative impact on the existing infrastructure.

"We don't do it that way" because it embodies too much risk is the response from someone who hasn't actually looked at all of the issues involved. And as far as it being a alternative with too much upfront expense, in many businesses, it's little more than the money that will be spent eventually, over time, solving the same set of problems. And how valuable is it to your business to have those problems solved now, rather than over the course of the next year?

In any case, only you and your team know your business well enough to make these types of choices. But don't let your embedded business practices have a negative effect on the way you update your datacenter. When considering the cost, make sure that business management understands the overall impact of major improvements; not just the upfront expenditure.

Related stories

• The 21st Century Data Center: An overview

• Is it better to own or outsource your data center?

• Open Compute: Does the data center have an open future?

• Own our own data center? Why would anyone do that, CIOs ask

• Datacenter optimization: Overhaul or incremental changes?

Today, the achievement seems amazing. More than 16 years without a glitch.  But the truth is more that the company had a use for the server for the last decade or so, not that the operating system kept working.  Because that was what NetWare did. It just worked. And while this story is amazing for the length of time the physical hardware survived, it’s only the most recent story about NetWare doing something that seems amazing, in a world where servers are reset, rebooted, and reconfigured at a rapid pace.

In 2001, a university found a NetWare server that had been lost for four years. IT knew it existed; they could see it and manage it on their network, but no one had any idea where it was physically located. It was eventually discovered during some renovations, when a hole was punched in a wall and it was found that a previous renovation had built a wall that blocked off the server closet where it was located.

Back in the NetWare days, servers staying up for years were not uncommon. Novell even had a page of screen shots from consoles showing extended uptimes that customers submitted. Domain controllers, file and print servers, application servers, it usually didn’t matter. The software just didn’t crash; the majority of users surveyed back then said that their servers were brought down only when they needed to be updated.

That’s the level of reliability that cloud service providers need to strive for. For users to depend on their services,  they need to just work. No excuses, no finger pointing, no questions. When clients boot up their systems, the provider services just need to be available. 

We’re getting there, but unplanned, unexpected, unprepared-for outages still plague the business, as well as a general unwillingness for providers to just say, “we screwed up; we will fix it and make it better.”

See also:

• Why velocity, not speed, is the key to cloud maturity
• Lesson learned from the Outlook.com outage
• Can a datacenter really drive economic recovery?
• Simple things make your datacenters more efficient

After all, if you are using a colo or cloud host for yourself, you must have already gone through the vetting process for that provider, otherwise why would you be using their services? And Client Connect is designed to give Savvis customers an enhanced ability to connect with other Savvis datacenter users in any of the Savvis datacenter facilities.

The Client Connect service, which Savvis customers need to voluntarily choose to participate in, both as buyers and sellers, is basically a directory of service providers hosting at any of Savvis' datacenter facilities. Each client gets a profile page where they can promote their service offering and where they need to provide contact information for potential customers. Offering your company's services in this fashion also includes the requirement for rapid deployment; Savvis is attempting to guarantee rapid client-to-client cross-connect provisioning, with a deployment completion target of within 48 hours when the service is ordered via the SavvisStation portal.

This is one of those perfectly obvious things that every large scale coclo and hosting services vendor should be providing. While a goal of using cloud services is to be able to not worry where the actual services reside, that's proven to be a somewhat elusive task. Weekly reports of even the largest service providers facing outages have to make customers think twice about where the service they contract for is being hosted. Especially given the degree of finger pointing that seems to occur when problems do crop up.

By promoting the other services running within the environment they control, Savvis would take the heat for any outages; giving their customers the proverbial "one throat to choke" for infrastructure issues. So making this service model available not only promotes customers of Savvis, but also shows them taking a stand for the reliability of their underlying infrastructure.

It's not much of a surprise that the datacenter business in Asia is booming. Not even including India, which is taking its own path into the world of the modern internet, Asia represents the fastest growing sector of connected users and businesses. This week's new datacenter openings reflect a good cross-section of the ways that datacenters are being used worldwide.

In Japan, Oracle announced the availability of a new datacenter designed to support the Oracle Right Now cloud service. Built completely on Oracle engineers systems the datacenter brings the Oracle RightNow Web Experience, RightNow Social Experience, RightNow Contact Center Experience, RightNow Engage, and RightNow Service Experience Platform to Asian customers from a local datacenter, which Oracle said will provide better control and management of the flow of data than running from a datacenter located on the other side of the world.

In China, Pacnet opened a 14,000 square-meter tier 3 datacenter facility that is carrier neutral and can support up to 2,000 high-density, power-hungry racks. The datacenter provides direct access to Pacnet's point of presence in China, as well as their wholly owned undersea cable infrastructure. The datacenter is targeted at enterprise and carrier customers doing business in China, but at this time, Pacnet has not disclosed any information on the tenants of the new facility.

In Singapore, Equinix announced that it would be adding a third datacenter in Singapore; a 385,000 square-foot facility that will support up to 5,000 racks. Targeted at financial and cloud services companies, the datacenter is scheduled to open in Q2 of 2014.