Thursday, May 9th, 2013

Time	Event
12:54p	Cloud Growth Slows at Rackspace, Which Cites OpenStack Transition The community around the OpenStack cloud platform is growing quickly, but sales are coming more slowly. That’s the story, at least, at Rackspace Hosting, the primary champion of the OpenStack platform. Rackspace said yesterday that its sales slowed in the first quarter, as customers are taking their time in adopting its new OpenStack cloud offering. Rackspace also faced headwinds in the face of pricing cuts from Amazon Web Services, the market leader in cloud computing, and a sluggish environment for enterprise IT spending. The sales slowdown follows Rackspace’s shift to an all OpenStack cloud late last year. Customers are excited about the potential of OpenStack, which is seeing huge growth in its developer community. But customers on Rackspace’s legacy cloud platform are taking their time in switching to the OpenStack cloud. As a result, the cloud revenue growth rate in the frist quarter slowed to 4 percent, down from 10 percent in the fourth quarter of 2012. Shares Poised to Drop Investors are not thrilled with the news. Shares of Rackspace (RAX) plunged 20 percent in off-hours trading, marking the second consecutive quarter in which Rackspace’s earnings report has led to a sharp decline in its share price. Rakspace stock fell 20 percent in February after the company said its cloud revenue growth has moderated. “The Q1 growth is not what we hoped for, but it’s also not outside the bounds of how we felt things could unfold given the new platform we are building,” said Rackspace CEO Lanham Napier in a conference call with analysts. “We completed the launch of our OpenStack cloud platform in November. And since that launch, we have seen no incremental growth on our legacy public cloud platform. This is largely because customers running on the legacy public cloud platform aren’t increasing activity.” Meanwhile, customers are running pilots on the OpenStack platform, but taking their time in deploying their apps on the new platform. “This product cycle transition will likely continue for some time given the large number of customers still using the legacy public cloud platform,” said Napier. Cloud Growth as a Differentiator The growth of cloud computing made Rackspace a hot stock in 2012, when the company’s shares rose 72 percent, following a 37 percent rise in 2011. As Rackspace has focused its business on the future of cloud computing, the growth of its cloud revenue has helped insulate the company from the variability of enterprise demand in the managed hosting sector, which has historically supplied the lion’s share of business for Rackspace. In a way, this quarter’s performance underscored the wisdom of the shift to a cloud model to “future proof” Rackspace as businesses change the way they provision IT infrastructure. Cloud growth slowed at Rackspace in the first quarter, but enterprise managed hosting slowed even more. “The place where we felt the growth reduction was in our enterprise business, which is primarily on our dedicated (server) form factor,” said Napier. “What’s happened to us here is we lost the momentum in our net upgrade activity in our enterprise segment. That was our fastest-growing segment in that chunk of the business, in our dedicated business historically.” The Amazon Factor Napier said the company remains deeply committed to OpenStack and believes that open source clouds – and specifically hybrid clouds – will ultimately drive growth in cloud adoption by businesses. Much of the early cloud adoption has been driven by developers, the target audience for Amazon Web Services (AWS).  Napier acknowledged that Amazon has a strong market position. “It’s clear that Amazon Web Services is doing really well,” said Napier. “AWS absolutely has a head start and has a lot of developer traction. Our personal belief is that open source is a better development model and framework than a proprietary system. Over time, there will be more developers contributing code to OpenStack than Amazon Web Services could ever hire. I just think right now, they got a developer head start, and it’s early in the game.” Then there’s pricing. Amazon continues to be aggressive in its pricing for cloud services. Although Rackspace offered some pricing breaks on cloud bandwidth and storage to support the OpenStack cloud launch, Napier said the company is following a different path than Amazon on pricing. “We obviously observe what others are doing in the market, but we are a cost-plus shop,” said Napier. “We price at a premium. We believe this premium reflects the fact that we have a differentiated service that we call Fanatical Support. So we will always be at a premium to other people out there because we’re the best service, we’re the best customer outcome. You’ve got to pay the most for that.” Napier said that even though revenue growth has moderated, Rackspace is making the right moves to compete in the cloud computing market. “We really feel like the OpenStack movement has hit a tipping point; that it basically has won in terms of freeing the open source standard,” he said. “So we’re going to keep our investment levels up, and we believe we have an incredible market opportunity. What we want to communicate to investors is right now we’re going for it.” (Comment on this)
1:30p	Cologix Now Even More Wired at the INFOMART Rob DeVita, the General Manager of Cologix Texas, in the company’s new data center space in the INFOMART in Dallas. (Photo: Rich Miller) DALLAS – The Cologix story is about connectivity, and begins at one of the country’s wired buildings – the Dallas INFOMART. The colocation company got its start in 2010 when it acquired a NaviSite data center on the first floor of the INFOMART, the iconic building on Stemmons Freeway that serves as a major intersection .for Internet traffic. “We’re going to go where the connectivity is, and that means the major carrier hotels in the central business districts,” said Rob DeVita, General Manager of Cologix Texas. “That’s really our model.” The company has followed that model as it has expanded, primarily through acquisitions, assembling a footprint in key network hubs in Minneapolis, Montreal, Toronto and Vancouver. Cologix now owns 12 data centers across these five markets, and sees a growth strategy in building connectivity ecosystems in second-tier markets. But it’s also thriving in Dallas, where it all started. Earlier this year Cologix opened a new data center on the second floor of the INFOMART, adding 12,000 square feet of space to the 28,000 square feet in its first-floor facility. The new space provides flexibility in supporting customers with a range of power density requirements. “The beauty of this facility is that we can have telco here (on the first floor) and then high-density clients upstairs,” said DeVita. Hot Aisle Containment To support those high-density clients, Cologix is using a data center design featuring hot-aisle containment, combined with in-row cooling products from APC by Schneider. When warm air from server exhaust enters the hit aisle, it is recirculated through the cooling units, which are spaced out along each row of cabinets, and then returned into the cold aisle. This approach provides flexibility in a multi-tenant environment where customers may have diverse IT gear residing in adjacent cabinets. “It allows us to put a 15 kW cabinet next to a 2 kW cabinet and not have any problems,” said DeVita, who says Cologix will use this approach in all of its new data center space. While the design requires some additional up-front investment, it provides lower operating costs on an ongoing basis, he said. Cologix is building a significant presence in Montreal, where it has six data centers and hosts the newly-opened Quebec Internet Exchange (QIX). “With access to eight million Quebec residents, and with 2 million households with fixed line broadband subscriptions, Montreal is the natural home for the new Quebec Internet Exchange,” said Sylvie LaPerriere, Chairperson of the QIX and Program Manager for Peering and Content Distribution at Google. “QIX represents a critical step towards creating a more robust domestic Internet, not just in Quebec, but across Canada.” Wired at the INFOMART Cologix hosts similar Internet exchanges in Toronto and Minneapolis. It also continues to add connectivity options in Dallas. Last month network service provider Hurricane Electric established a point of presence (PoP) within the Cologix facilities at the INFOMART. Hurricane operates the world’s largest native IPv6 network, which means its network will increase in importance as the shift to IPv6 picks up speed. The INFOMART has more than 30 carriers in the building, which was built by Trammell Crow in 1985 as a trade center for the IT industry. Its stunning design, which features a glass facade and seven-story atrium, was modeled after London’s Crystal Palace, the site of the first World’s Fair. In the late 1990s the INFOMART emerged as a primary “carrier hotel” housing telecom firms and data center providers. The building is now owned by DCI Technology Holdings. “It’s the most fiber-dense building you’ll see in the South Central U.S.,” said DeVita. But that also means that Cologix isn’t alone in offering colocation space at the INFOMART, where the tenant list includes ViaWest, SoftLayer, and colo market leader Equinix. But DeVita says Cologix has found an opportunity in the wired nature of the INFOMART. As the building’s roster of ISP and data center tenants has grown, not all of the cabling has been seamless. “We’ve wired the building to untangle the mess of trying to do in-suite connections in the INFOMART,” said DeVita. “We’ve got multiple diverse routes to all of those suites.” “People are going to come to the Infomart for the network providers that are here,” he added. “The value of the Infomart and its connectivity is the reason people want to be here. We believe the value proposition is pretty strong.” A row of cabinets in the new Cologix data center at the Dallas INFOMART. (Photo: Rich Miller) (Comment on this)
2:30p	Why Data Centers are Shifting Towards Converged Infrastructures With more virtualization, IT consumerization and a lot more data – the data center needs a new way of deploying high-density computing platforms. In this white paper, IDC (with sponsorship from VCE) outlines the rapid growth in high-density computing needs. By downloading this paper, you can learn how IDC  found that higher utilization of IT assets and operational efficiency — which results from running more virtual machines on new-generation servers — reaches a plateau and often levels out at a certain point. This happens because the shift to virtualized servers often leads to strains in other areas of the infrastructure: Virtual server sprawl increases server/storage/network stress and the accompanying administrative burdens required to deal with this stress. This makes support/maintenance more challenging and threatens application performance. Handling this anticipated pressure by overloading/overprovisioning storage and data network facilities forces time-consuming, costly, and often unnecessary hardware upgrades. Application performance and recovery behaviors (data recovery and cleanup) on error conditions can vary unexpectedly, stalling plans to migrate more business critical applications to virtual environments. The reality within the data center world is pretty direct: The infrastructure management challenge in has increased, as depicted in the “Virtualization Management Gap” in Figure 1. Furthermore, IDC and this white paper indicate how recent research shows that IT departments now spend three-quarters (76.8%) of their time and resources maintaining the environment and less than a quarter (23.2%) on value-added activities. In this white paper, IDC outlines how the data center is entering a new business cycle in IT where customers are prepared to trade choice for both ease of installation and simplicity of management. While the majority of customers are still evaluating converged systems, over a quarter (27.4% — approximately double the previous year’s percentage) are currently using or planning to use them. IDC expects adoption to increase as 44% (also double the previous year’s percentage) of those considering convergence will likely adopt in the next three years. There are direct data center technologies which help deliver high-density computing, excellent resource utilization and of course the ability to converge an infrastructure. Formed by Cisco and EMC with investments from VMware and Intel, this white paper outlines the vBlock platform and how VCE markets an integrated converged infrastructure solution for the data center. In fact, VCE develops a range of platforms and solutions for virtualized environments based on components from Cisco, EMC, and VMware. Download this white paper today to learn how VCE’s goal is to accelerate the adoption of converged infrastructure and cloud-based computing models which prove to dramatically reduce the cost of IT while improving management and growth capabilities. (Comment on this)
3:49p	Cray Extends XC30 to the Technical Enterprise Cray announced a new addition to its XC30 systems, with the XC30-AC (air-cooled) supercomputer. The new model includes all of the advanced high performance computing (HPC) technologies offered in the Cray XC30 system, and features aggressive price points intended to attract a new a class of HPC users — the technical enterprise. The air cooled model economizes the packaging, networking, cooling and power options of the Cray XC30, while balancing customer price/performance requirements.  Prices for the XC30-AC start at $500,000. The XC30-AC enables technical enterprise HPC users in a variety of industries to apply supercomputing resources towards solving technical and business challenges. Additionally Cray has a line of cluster supercomputer offerings, the CS300 cluster supercomputers, to compliment the XC30-AC systems. “Innovation is not limited to Fortune 100 companies. There are many Fortune 1000 companies, and even departments within Fortune 100 companies, with a growing need for a supercomputing system that provides a critical tool for taking advantage of performing complex simulations,” said Peg Williams, Cray’s senior vice president of high performance computing systems. Cray states that the new model has all of the features and functionality of its high-end Cray XC30 systems, and is suited for technical enterprise customers, helping them access supercomputing capacity at a lower price. The XC30-AC has allowed Cray to enter new markets for its supercomputers, with early customers including a global consumer electronics company and a global financial services company. It is targeted at less-traditional HPC segments, as well as a broader class of users in more traditional HPC markets, such as academia, defense and earth sciences. While the XC30-AC contains many of the features of the XC30, it differs in that it is air cooled, has physically smaller compute cabinets with 16 vertical blades per cabinet, has a single fan for bottom-to-top vertical air flow, and has a lower power option supporting either 480V or 208V operation. “Cray has a history of leveraging the supercomputing technologies featured in their high-end systems, and economically repackaging those same technologies to offer solutions to fit the needs of HPC users with smaller budgets,” said Earl Joseph, IDC program vice president for HPC. “Simulation is no longer bound by the high-end data center, and Cray’s new XC30-AC system continues the company’s tradition of creating purpose-built systems that appeal to new customers in expanding segments of the supercomputing market.” (Comment on this)
4:05p	Design for a Dynamic Environment This is the forth article in a series on DCK Executive Guide to Data Center Designs. Historically, data center IT loads have been relatively stable and predictable if viewed over a 24 hour or weekly period. This is beginning to change for several reasons. The first is virtualization, which originally allowed for individual applications which were running on distributed and underutilized servers, to be consolidated on to more centralized hardware resources such as bladeservers, resulting in higher CPU and overall server utilization, contained in less space. More advanced virtualization software offers energy management features which can monitor computing demands. Excess resource capacity such as un-utilized servers can be put into low power sleep modes or even be powered off automatically when not needed, but which would power up and then be put back on-line as computing demands rise. The second reason is that the IT hardware itself became dynamic while becoming more energy efficient. Instead of wasting substantial amount of power when idle, they now reduce power significantly when idle, yet draw more power (and generated more heat) when called upon to do work. The US EPA Energy Star program for data center equipment requires this for Energy Star certification of IT equipment such as servers, since 2009 and now is in the process of finalizing the standards for Storage and Network equipment. (see part 3 Energy Efficiency) The result is twofold; the overall total IT power and cooling load has begun to vary more over time as the amount of computing load increased and decreased over a 24 hour cycle. Moreover, the heat IT loads have begun to shift from rack -to-rack and row-to-row, in response to demand driven computing activity, creating traveling hot-spots across the data center. While the overall goal is to improve the energy efficiency of the IT systems, this has challenged a lot of older more traditional cooling system which were not designed to handle these new more dynamic conditions. When considering a new data center design, the IT team needs to work with the facility design team to provide more information on the type of hardware they plan on using, as well as any of the energy management features of the virtualization software, which can impact the design of the cooling system. The complete Data Center Knowledge Executive Guide on Data Center Design is available in PDF complements of Digital Realty. Click here to download. (Comment on this)
4:30p	Big Switch Networks and F5 Partner to Simplify Network Orchestration Big Switch Networks announced a joint solution with F5 to simplify how private cloud data center networks are built and orchestrated, enabling customers to deploy the rich network services they expect from traditional networks in an automated fashion consistent with the provisioning speeds of virtual workloads. Using the Big Virtual Switch network virtualization application, it provides two-way coordination between virtual network segments and F5 Application Delivery Networking services. The combined solution integrates the Big Virtual Switch network virtualization application, BIG-IP Application Delivery Networking (ADN) services and OpenStack into a flexible and unified cloud orchestration framework, using the F5 BIG-IQ Cloud REST API. Big Virtual Switch dynamically provisions virtual network segments, or slices of the network, for applications and workloads and enables automation. The solutions supports the BIG-IP iControl API and the BIG-IQ Cloud REST API. It will also support both native management and OpenStack Horizon Management Dashboard. “Inserting ADN services, and doing so automatically, is very important in virtualized network and SDN environments, and can only be achieved with a truly programmable system,” said Mansour Karam, Big Switch Networks vice president of business development and strategic alliances. “Big Virtual Switch and BIG-IQ from F5 are both prime examples of programmable systems, and we are enthusiastic about working together to deliver a network automation solution.” “Intelligent networking continues to be at the forefront of customers’ minds,” said Jason Needham, VP of Product Management and Marketing at F5. “Integrating application delivery components with emerging SDN and cloud architectures provides a path for organizations to seamlessly increase efficiency and add programmability to their systems. As a key innovator in the SDN space, Big Switch is a great partner, and we look forward to adding further value for joint customers through enhanced efforts with their team.” (Comment on this)
5:10p	Interop News: Mellanox, Marvell, Arista Networks Interop 2013 is being held this week in Las Vegas, with announcements coming from Mellanox, Marvell, and Arista Networks.  The Interop event conversation can be followed on Twitter hashtag #interop. Arista named grand winner for Best of Interop Arista Networks announced it  was named the Best of Interop 2013 winner and the Arista 7500E Modular Data Center Ethernet Switch has won both the Grand Prize and the Networking category, recognizing innovation and technological advancements. “Arista is honored to be recognized again for delivering the best products in the networking industry. The reception on the Arista 7500E from analysts and customers is beyond our wildest expectations,” said Jayshree Ullal president and CEO for Arista Networks. Eric Hanselman, chief analyst at 451 Research noted, “The 7500E is the winner in both the Networking and Grand Award categories, and its impressive specs are part of the reason. One of the key pieces of innovation in the 7500E is the embedded optical capability that enables transitions between today’s and tomorrow’s networking needs at a reasonable cost.” Mellanox announces Ethernet Interconnect and InfiniBand/Ethernet gateway Mellanox Technologies (MLNX) announced its end-to-end 56 Gigabit Ethernet product line, an Ethernet-based server and storage interconnect solution. The 56GbE solution, consisting of Mellanox’s  ConnectX-3 and ConnectX-3 Pro NICs, SwitchX-2 based SX1024 and SX1036 switches, QSFP+ cables, and acceleration and management software, delivers 40 percent more bandwidth than competing 40GbE solutions, enables higher server density at lower cost, as well as a more efficient network design using less cables and less switches. “Increasing bandwidth requirements, data volume, data center consolidation and virtualization are driving the demand for faster Ethernet solutions,” said Gilad Shainer, vice president of marketing at Mellanox Technologies. “Utilizing Mellanox’s complete 56GbE solution, companies can gain significant application performance by connecting their storage and server infrastructure with the fastest Ethernet fabric. Furthermore, 56GbE enables the most efficient inter-switch connectivity resulting in lower overall data center capital and operating expenses.” Mellanox also announced the next step in its Virtual Protocol Interconnect (VPI) strategy, with the availability of an integrated InfiniBand to Ethernet gateway in its SwitchX-based InfiniBand and Ethernet switch systems. Using this new gateway, Mellanox edge switches can now be used for both InfiniBand and Ethernet switching at the same time with an integrated gateway capability between the two networks. “Mellanox’s new InfiniBand to Ethernet gateway functionality built within Mellanox switches provides the most cost-effective, high-performance solution for data center unified connectivity solutions,” said Gilad Shainer, vice president of marketing at Mellanox Technologies. “Mellanox’s systems enable data centers to operate at 56Gb/s network speeds while seamlessly connecting to 1, 10 and 40 Gigabit Ethernet networks. Existing LAN infrastructures and management practices can be preserved, easing deployment and providing significant return-on-investment.” Marvell announces ARMADA 375 System-on-Chip Marvell (MRVL) introduced the Marvell ARMADA 375 System-on-Chip, a dual-core Cortex A9 SoC platform that builds on the highly successful ARMADA 370 and ARMADA XP families of embedded ARM processors for business networking applications. The ARMADA 375 SoC, available in 800 MHz and 1 GHz speed, incorporates a host of I/O peripherals, and is optimized to consume very low power for a wide range of applications. It is designed to integrate a 32-bit DDR3/3L memory controller, a security engine, SATA 2.0 ports and dual PCI-Express interfaces to make system designs simple and economical. “With the launch of the ARMADA 375, Marvell is setting new standards for SoC system performance and integration at very low power for a dual-core 1 GHz processor,” said Ramesh Sivakolundu, vice president and general manager, Connectivity, Servers, Infrastructure Business Unit (CSIBU) at Marvell Semiconductor, Inc. “We are enabling our customers to address a wide range of performance and cost-sensitive application challenges by leveraging their existing investment in the legacy Marvell ARMADA SoC families.” (Comment on this)

<< Previous Day

2013/05/09 [Calendar]

Next Day >>