Wednesday, September 10th, 2014

Time	Event
12:30p	Amsterdam Internet Exchange Moves Into San Francisco Data Center After establishing presence on the east coast, AMS-IX USA, subsidiary of AMS-IX (Amsterdam Internet Exchange), is expanding to San Francisco. The company has established a point of presence at Digital Realty Trust’s San Francisco data center at 365 Main Street. Called AMS-IX Bay Area, the peering exchange builds on progress made with AMS-IX in New York. AMS-IX and a number of other European internet exchange operators, including LINX (London Internet Exchange) and DE-CIX (Deutscher Commercial Internet Exchange) recently established presence in data centers in New York and Northern Virginia markets in parallel with establishment and development of Open-IX, a non-profit industry organization chartered with increasing the number of Internet exchange options in the U.S. Open-IX has devised two sets of certification requirements for exchanges and data centers that host them. Digital’s 365 Main facility has been certified, and AMS-IX Bay Area is “compliant” with Open-IX standards, according to AMS-IX. Together with Open-IX, European operators have brought what is often referred to as the European exchange model to the U.S. The European exchanges are member-governed and distributed, while the leading U.S. exchanges are inside a handful of commercial providers’ data centers, controlled by the providers, primarily Equinix. Digital Realty takes part in the Open-IX effort, and so do its competitors, such as DuPont Fabros Technology, CyrusOne and RagingWire, all in hopes that hosting exchanges will make their facilities more attractive for customers. “The expansion of the AMS-IX/Digital Realty relationship from New York to San Francisco further strengthens and expands Digital Realty’s commitment to the Open-IX initiative,” John Sarkis, general manager of colocation and connectivity at Digital, said. “We look forward to continuing to provide our clients with the widest possible range of connectivity options.” On the east coast, AMS-IX has agreements with Digital Realty for a point of presence at Google-owned 111 8th Avenue, with DuPont Fabros for infrastructure in its Piscataway, New Jersey, data center, with Sabey Data Centers for presence in its Intergate.Manhattan building and with owners of the 325 Hudson carrier hotel, also in Manhattan. As of Tuesday evening, seven entities were connected to AMS-IX New York, driving peak traffic at 35.91 Mbps. Together with the exchange in Amsterdam, the organization has about 700 connected parties with peak traffic rate approaching 3 Tbps. (Comment on this)
12:45p	Why Your IT Fails in the Summer, Overcoming Outages John Gentry is the vice president of marketing and alliances at Virtual Instruments. There are plenty of perks about the summer months; unfortunately, high-functioning IT infrastructure doesn’t seem to be one of them. Poor performances and full-blown IT outages occur most commonly June through August. The U.S. State Department, for example, suffered database performance issues last month that resulted in tens of thousands of travelers being stuck overseas, and Facebook has experienced a few significant outages since June. So why is this? There are a few causes to blame. The perfect storm for IT in the summer First and foremost, there’s the issue of vacation time. Warmer weather and longer days make for perfect vacation conditions. Around 44 percent of American employees take some sort of time off from June through August. While this doesn’t always mean it’s really time off from work – working remotely and choosing flexible hours are becoming increasingly popular options – it does mean time off from the physical office building, which means there’s a skeleton IT crew in place when disaster strikes. There’s also the fact that many businesses decide the generally slower summer months are a good time to take on the task of moving their workloads to the cloud and adopting more complex virtualization strategies. This causes strain on an already under-managed infrastructure, making problems with system performance even more likely. Lastly, spikes in travel during these months lead to spikes in cellular and Internet activity, which can overwhelm an infrastructure that’s struggling to operate normally as is. So what can businesses do to avoid these summer outages? There is unfortunately no magic solution that will sufficiently mitigate each of these issues, but there are a few tips and tricks to ensure your data centers remain up and running at their maximum potential, even during this risky season. How to avoid the meltdown Plan ahead and get your disaster recovery in place. The hope-and-pray method is a risky move. Instead, consider the very real possibility that you will experience summer outages of some kind, and then prepare accordingly. Whether a hurricane hits or your IT guy hits the beach, having recovery options in place for your mission-critical workloads is crucial. Just make sure you have the proper IT performance management and monitoring ready to keep this backup running smoothly. Put your mission-critical workloads where you can see them. I don’t mean setting up your desk in the same room as your servers, but you should keep those workloads on your servers and not in the public cloud. While the public cloud certainly offers some benefits, in-depth visibility to infrastructure performance isn’t one of them. Without those kinds of granular metrics, it’s impossible to know what you’re getting for your cloud consumption and purchases, and when it comes to your mission-critical activity in the sensitive summer months, that uncertainty is unacceptable. Create a private cloud for your infrastructure performance monitoring needs. With the proper infrastructure performance management tools in place, your IT team can get a comprehensive assessment of the entire private cloud infrastructure, including real-time information and data for proactive choices about utilization, performance and existing capacity for each piece of the cloud’s puzzle. This leads to smarter decisions and ultimately greater infrastructure reliability. These are processes every IT team should put in place to sidestep these summer challenges, but they offer valuable food for thought for the rest of the year, as well. Is your business’ infrastructure prepared for any obstacles that may come its way? Summer presents a higher risk of outages, but you never know what’s around the corner in any season. Make sure your IT team and the equipment they manage are well equipped at all times, and your business will be ready to run smoothly 24/7/365. Industry Perspectives is a content channel at Data Center Knowledge highlighting thought leadership in the data center arena. See our guidelines and submission process for information on participating. View previously published Industry Perspectives in our Knowledge Library. (Comment on this)
3:48p	Digital Realty Sells 80% of Ashburn Data Center to Griffin Capital Digital Realty Trust announced a $188 million joint venture with Griffin Capital Essential Asset REIT. Digital Realty is contributing a 130,000-square-foot turn-key data center in Ashburn, Virginia, valued at about $185.5 million, excluding $2.1 million of closing costs, or $20,611 per kW. The joint venture has arranged a $102 million five-year bank loan. GCEAR will hold an 80 percent interest in the joint venture and Digital Realty will retain the rest and continue to manage the property for management fees. The transaction generated net proceeds of about $168.4 million for Digital Realty, which will help the REIT pay down debt and fund other investments. The deal gives GCEA a fully leased property expected to generate about $13 million per year in revenue. The average remaining tenant lease term is about seven years. The venture is part of Digital Realty’s plan to divest non-core properties in response to changing market dynamics. The company is getting rid of the least-performing properties in its massive global portfolio and changing its mix of product offerings to include solutions that combine its traditional space-and-power offerings with services higher up the stack provided by some of its customers. Jeff Hoppen, managing director of capital markets for Digital Realty, said, “This joint venture transaction is a significant milestone for Digital Realty, as it furthers our objective of maximizing the menu of available capital options, while minimizing the related cost. The transaction also has the ancillary benefit of reducing our tenant concentration while establishing an attractive private market valuation benchmark for our Turn-Key Flex properties.” The company formed a similar joint venture with Prudential Real Estate Investors on a fully leased 110,000-square-foot facility in Somerset, New Jersey, in October 2013. It also recently sold a property to an unnamed single tenant that was occupying it for $42 million. (Comment on this)
4:06p	365 Data Centers Raises $16M Series B, Closes $55M Credit Facility 365 Data Centers has raised $16 million in a Series B funding round and secured a $55 million credit facility. The announcement comes on the heels of adding a new management team and changing corporate identity. The company said it will use the funds to grow its colocation business in top and second-tier U.S. markets, develop new data center products and expand cloud and managed services offerings. 365 Data Centers’ namesake comes from the iconic 365 Main facility in San Francisco, now owned by Digital Realty Trust. In 2012, 365 Main co-founders Chris Dolan and James McGrath resurrected the brand by acquiring 16 data centers from Equinix. Equinix was looking to divest some facilities it got through the Switch & Data acquisition. 365 acquired the facilities, seeing an opportunity to create a “national player with a local focus.” Focusing on small and medium businesses, it ditched the traditional multi-year colo contract, offering services on a month-to-month basis to better appeal to its target customers. The company said it achieved triple-digit sales growth year over year and broke its all-time quarterly sales record by more than 20 percent. John Scanlon, a 30-year industry veteran, joined as CEO earlier this year. “365 Data Centers brings greater flexibility and accessibility to clients through solutions that simplify colocation services to bring them more in line with cloud consumption models,” said Scanlon. “This new financing allows us to continue executing on our growth strategy, whether organically or by acquisition. “We’ve put together an experienced management team, created a strong growth plan and delivered great results in a matter of months,” he continued. “We have the faith and support of capital markets behind us, and when talent meets capital, good things happen.” The Series B includes participation from existing investors Crosslink Capital and Housatonic partners. Gary Hromadko, partner at Crosslink, said, “With the largest cloud market in the world and increasing cloud adoption driving data center and colocation services expansion, North America represents a great market opportunity with expected growth of approximately 15 percent this year to more than $8 billion.” The credit facility came from Fortress Credit Corp., a Fortress Investment group affiliate. Aaron Blanchette, managing director at Fortress Investment Group, said, “In our view, 365 Data Centers is similarly positioned to succeed in the expanding data center and cloud industries. We look forward to playing a part in 365’s organic expansion and potential mergers and acquisitions, as opportunities arise.” (Comment on this)
5:28p	Another QTS Exec Joins Canara Canara, a recently rebranded company that provides monitoring and predictive analytics for data center backup power systems, continues to add to its executive team. Helmed by former QTS Realty Trust executive Tom Mertz, the company has added another QTS vet. Chris Oberkfell, formerly vice president of sales operations at QTS, has joined Canara as the chief operating officer. The addition is the latest step in executive team changes Mertz has been making since he took the CEO role in June. In July, Mertz brought on board chief business officer Steve Manos, well-known in the industry from his time at Lee Technologies, a data center management services company acquired by Schneider in 2011, and from his frequent speaking engagements. “Chris has a broad set of experiences in the data center industry and ability to adapt, manage and improve processes to drive business performance,” Mertz said about Oberkfell. “With his knowledge of the industry and ability to help organize and develop the team, he is an important addition to the Canara leadership team.” Batteries are often at fault in data center outages when they don’t kick in as expected. The San Rafael, California-based company’s devices continuously analyze data center UPS batteries to predict when they are going to die, helping avoid unexpected failure. The value proposition is risk mitigation, higher efficiency and more reliable backup power systems. Canara (formerly IntelliBatt) recently raised $4.25 million in equity from Columbia Capital in support of an aggressive growth plan. Mertz came on with a vision for expanding the leadership. Canara’s customers include large companies such as Barclay’s, CenturyLink, Cologix, Comcast, Digital Realty Trust, Equinix, FedEx, PG&E and Sony. (Comment on this)
5:50p	No “Friday Favors” for Missouri Data Centers Missouri Democratic Governor Jay Nixon is criticizing several sales tax exemptions targeted to data centers. The Examiner reports Nixon is particularly focused on 10 bills he says would cost local governments $351.2 million annually, as well as $425.1 million lost to the state. Legislators passed the 10 bills on the final day of the session, all of which offered tax breaks to businesses. Nixon called the bills “Friday favors,” a commentary on last-minute lobbyist appeasement, and vetoed all 10. One of those bills is designed to encourage the development of data centers and accounts for two-thirds of those projected local losses of revenue. Both sides are unlikely to override Nixon on the bill, despite the Missouri General Assembly being heavily Republican. The argument for vetoing is runaway tax credit programs in the state hampering the ability for it to provide in other areas. The argument for the bill is data centers are viewed as extremely positive for local communities, and several states are offering increasingly aggressive tax breaks to attract the business making for a competitive landscape. Earlier this week, an economic development department in a small Ohio town, for example, recommended that the city give nearly $7 million worth of city-owned land to Amazon free of charge, in exchange for a commitment to build a 750,000-square-foot data center complex there. The Missouri bill in question is Senate Bill 584. It lacks incentive conditions like minimum job creation and capital investment requirements. The bill would qualify all existing facilities regardless of whether they’re expanding. For the sake of comparison, the proposed land incentive in Ohio is tied to commitments to build within a defined timeline. The bill is unlikely to make it past the senate and has slim chances to get by the house, even if it makes it that far. There’s a concern that it is so broadly written that businesses outside of data centers can take advantage, costing even more in revenue. The veto comes as no surprise, Nixon citing his intention to veto the bill after it initially passed on May 16. The argument from Nixon is that it lacks in taxpayer protections, something included in earlier versions. “On the last day of the session, without a fiscal note, without a hearing, they added an amendment that is going to cost hundreds of millions of dollars and that doesn’t have any of the taxpayer protections that were in the data center legislation that was moved forward earlier on,” Nixon told Kansas City Business Journal. Nixon vetoed a record 33 bills in total, the most by any Missouri governor since 1961. A governor in the midst of the Ferguson police-shooting controversy, he has also vetoed several state budget line items. (Comment on this)
6:00p	Open Compute Hardware Not Cheap Enough for Enterprises Open source “vanity-free” hardware bought in bulk from Taiwanese manufacturers may offer a compelling price difference for the scale of Facebook, but people who work in the majority of the world’s enterprise IT shops generally don’t view servers promoted by Facebook’s Open Compute Project as something that makes sense for their data centers. Yesterday we wrote about the role of OCP in the enterprise data center as seen from OCP hardware vendors’ and the Open Compute Foundation’s perspectives. Today, we’ll cover opinions of data center industry experts who work with enterprise data center end users. Come see a Data Center Knowledge panel on Open Compute at the Data Center World Orlando conference in October Users that don’t operate hyperscale data centers and don’t make hardware purchases of corresponding size either don’t see a price difference compelling enough to make the switch or are simply not aware of the alternative ways of looking at hardware purchasing. It all comes back to scale “The old adage, ‘no-one ever got fired for buying IBM/HP/Dell’ still holds for most of those folks,” Mark Monroe, CTO and vice president at DLB Associates, wrote in an email. DLB is a consulting engineering firm that designs data centers, among other projects. Before joining DLB, Monroe was executive director of The Green Grid and before that led corporate sustainability strategy at Sun Microsystems. Monroe said that from what he’s heard, the price difference for OCP gear is not big enough to overcome inertia of business-as-usual and “risk-avoidance bias” most corporate IT managers have. When he worked for a “major systems vendor,” it was not uncommon for Fortune 200 firms to get up to 35 percent discounts from list price for large $10 million to $20 million annual purchase agreements. With discounts like that, even if OCP hardware is 40 percent cheaper than Dell hardware, taking the risk (real or perceived) is not worth it. Monroe warned that the numbers were not exact, “but they’re in the ball park.” Such purchase agreements would be for telcos or financial services companies with comparatively large IT portfolios. A lot of the companies “clustered around number 200” on the Fortune 500 list don’t buy anywhere near that amount of hardware. Companies like Starbucks or Kellogg’s cannot absorb $10 million worth of hardware per year. “There’s probably 100 companies in the world that can absorb more than that,” Monroe said. “Even eBay, at around 10,000-20,000 servers per year considers themselves pretty small compared to the big hyperscale folks.” OCP visibility in enterprise market minimal KC Mares, who’s done everything from consulting Equinix on energy to running global data center strategy at Yahoo, pointed out that OCP was also too new for many IT organizations. If an IT manager doesn’t go to industry conferences or read IT press often, they mostly hear about IT from those who sell to them, companies like Dell, HP, Cisco, EMC and NetApp. “And while OCP is talked about in data center press and conferences, it’s still new to most IT folks and they are afraid to stick their neck out on a purchase with someone other than a well-known brand,” Mares wrote in an email. “Most IT managers don’t really care that Facebook endorses it, as they are not Facebook and have different needs.” Lack of awareness is not to be underestimated. There’s a popular belief out there, for example, that OCP hardware vendors don’t offer support, which makes it even harder for an IT manager to switch from an all-inclusive agreement with an incumbent vendor. Frank Frankovsky, president and chairman of the board of the Open Compute Foundation, said this was not true and that OCP vendors offered multiple levels of support. But that’s part of the foundation’s job: to disseminate information and separate fact from fiction. At least one OCP vendor we talked to did agree that the hardware was not really much cheaper than incumbent gear upfront. Hyve Solutions president Steve Ichinaga said cost benefits really kicked in overtime, expressed through operational costs. (Comment on this)

<< Previous Day

2014/09/10 [Calendar]

Next Day >>