[Most Recent Entries] [Calendar View]
Tuesday, March 17th, 2015
- Hardware and Software – design and monitoring
- Operations – operational runbooks and root cause analysis updates
- Operating Environment (DCIM) – data center environment
- Security – malware vigilance
- Disaster Recovery – remote/local choice depending on natural conditions (floods, earthquakes, tsunami, etc.)
| Time | Event |
| 12:44a | NetDepot Data Center Outage Causes Prolonged Downtime for Customers A data center outage in Atlanta that started on Sunday evening brought down servers used by many customers in a prolonged period of downtime, which according to some of them, lasted for more than 20 hours. Customers of NetDepot, the data center service provider that operates the Atlanta facility, took to Twitter to vent their frustrations with the protracted outage and with what appeared to have been lackluster communication by the provider during the incident. Joshua Distler, founder of LiveSurface, an online library of image templates for designers, told us via Twitter that servers that hosted the LiveSurface website in the data center went down around 9:30 p.m. Eastern on Sunday. They remained down as of Monday evening, he said. According to Distler, the outage came during a scheduled maintenance period. NetDepot representatives did not respond to a request for comment about the data center outage in time for publication. The company said on its official Twitter feed Monday that there was a “network outage in one of our Atlanta zones.” The problem appears to have been caused by faulty VLANs (Virtual Local Area Networks), according to another Tweet by NetDepot: The company owns and operates its 75,000-square-foot Atlanta data center. Besides Atlanta, it has data centers in Dallas and Los Angeles. Many NetDepot customers complained on Twitter about a lack of information about the outage from the provider. Some complained about more than 15 hours of downtime; others said their servers were down for 20-plus hours. Here are some examples: |
| 12:00p | Equinix Rolls Out Global Channel Partner Program Equinix has launched a new global channel partner program to capture more revenue in the enterprise hybrid cloud market. In addition to a formalization of partnership activities, the company has expanded the types of partnership engagements it makes and increased the amount of resources behind its partnership initiatives. The program extends across the Americas, Europe, Middle East, Africa, and Asia Pacific. Equinix is increasing staff in support of the program 10-fold. It is expanding its global solutions architect team and will do more joint selling engagements with partners. Managed service providers, network service providers, system integrators and solution providers can combine with Equinix to design and deploy cloud and IT solutions for enterprise customers. These cloud channel partners can act as resellers, referral agents, or cloud technology platform partners. “What we are trying to do is enable customers and partners to get to the cloud,” said Pete Hayes, chief sales officer at Equinix. “This is high-level engagement around enterprise customer engagements.” The company’s increased focus on cloud channel partnerships is in response to growing enterprise need for IT infrastructure evolution to hybrid cloud. Customers require more than just colocation or cloud; they often require a mix as well as guidance. To help with that evolution, data center providers have been building their own diverse services, buying companies that already provide them, or partner with existing servce providers. Equinix’s approach has always been to take the partnership route. “In our data centers, network service providers and cloud service providers are going to be the motivating factors to get the enterprise customers to move workloads out of the basement,” said Hayes. Equinix has several partnerships already in place, but the program changes the dynamic of these partnerships. “In the past, Equinix had more referral types of partners,” said Christopher Rajiah, vice president of Worldwide Channel Sales and Alliances. “This expands to support different sales motions, resale, or technology partnerships. We’re taking different ways to partner and taking it to a global level. Channel is a key strategy going forward.” Partnering allows service providers to add “next-generation” cloud interconnection solutions to their existing product offerings. Channel partners can also tap into Equinix’s established ecosystems and products, such as Cloud Exchange and Performance Hub. The Cloud Exchange provides direct access to multiple clouds and multiple networks from a single port. Channel partners can also create new recurring revenue streams by implementing the Equinix Performance Hub, a solution that distributes a company’s data center infrastructure across multiple locations to accelerate application delivery to employees, customers, and partners worldwide. “Enterprise customers are trying to serve different communities and running into complications in globally supporting a dispersed customer base,” said Hayes. “There’s an enterprise need to deliver mission critical applications in cloud. This [program] is high-level engagement around enterprise customer engagements.” A similar rationale of enterprise IT evolution was behind the acquisition of cloud professional services firm Nimbo. It’s no longer a matter of enterprises wanting to move into an Equinix facility, but a matter of how. One hundred companies – enterprises, networks and cloud service providers – have joined Equinix Cloud Exchange. Along with top cloud platform providers, other service providers that have recently joined the Equinix Channel Partner Program include Carpathia, Datalink, Datapipe, Dimension Data, and Unitas Global. “Leveraging the AWS platform and the Equinix global data center network has allowed us to expand the range and reach of our managed enterprise hybrid IT solutions”, John Lady, Datapipe CTO, said via email. “The partnership with Equinix has been invaluable as we grow and expand the suite of managed services and geographic locations through which we provide services to our customers, and we look forward to continuing to work together to help our clients future proof their IT.” |
| 3:30p | The Evolution of High Availability Dr. Terry Critchley is a retired IT consultant and author who’s worked for IBM, Oracle and Sun Microsystems. Surely, I hear you say, availability requirements do not change except for striving to achieve higher availability targets through quality hardware and software. In true British pantomime tradition I would say, “Oh yes they do!” And I’m guessing you would like me to prove it! The traditional view of designing for and achieving high availability systems has been concentrated on hardware and software, although recently people have recognized the importance of ‘fat finger’ trouble causing outages. This is attributed to people (liveware) errors which cause outages or incorrect operations which can be construed as taking the system out. Such errors encompass such outages as entering the wrong time of day or running the jobs in the wrong order. This latter type of finger trouble brought down the London Stock Exchange on March 1, 2000. Although the reason is unspecified I suspect a leap year issue between the two systems – one recognized February 29, the other didn’t. If we can minimize the hardware and software issues and reduce finger trouble by rigorous operations procedures (leading to autonomic computing) then the problem is solved. Or is it? There are other factors which are either not recognized, understood or even thought of. Availability is usually defined as:
where full availability is represented by 1 or 100 percent. In addition, non-availability (N) = (1 – availability) x 100 percent, which is often expressed as a time (seconds etc.).
This is often a myth since the restoration of a failing piece of hardware or software does not mean the affected service is available at the same time. There are often other recoveries required, such as reconstitution of RAID arrays or recovery of a database to some predefined state. This latter recovery I call ramp up time which is additional to the time taken to fix/recover failing hardware and/or software. This can be less than, equal to or much greater than the hardware/software fix times. This equates to the fact that non-availability is really expressed by the equation:
This is a point often overlooked (deliberately or unintentionally) when stating availability percentages and times. A failure with a two minute fix time but a ramp up time of 120 minutes will blow any desired or offered 99.99 percent availability out of the window for several years. One example is a U.S. retail company which suffered an outage of its system a year or so ago. The problem was fixed in about an hour but they estimated that recovering the database to its original state would take several hours. This correct working state of a service is normally dictated and specified in a Service Level Agreement (SLA) between IT and the business. The business user is not really interested in the use of a duplicate network interface card to solve a server problem; he or she is interested in getting the service back as it was before the failure. The next and possibly newest outage generator are security leaks in the broadest sense. Most publications on high availability concentrate on the availability of the system whereas the availability of the business service is key. Recently, Sony estimated the damage caused by the massive cyberattack against Sony Pictures Entertainment was $35 million. TeamQuest developed a Five-Stage Maturity Model which they have applied to service optimization and capacity planning, along the lines of the older Nolan curve of the 1970s. In this model they specify, among other things, the view taken by an organization in each of the five stages. These are summarized in the table below.
Many availability strategies lie in the Reactive column, dealing with components; some are in the Proactive column where operations and finger trouble are recognized. I believe that organizations ought to migrate to the Service column and then the Value column. The key areas to design, operate and monitor in moving up the stages are: The viewpoints of interest are the end user of the IT service and the IT person’s view of the system. A computer system can be working perfectly yet to the end user be perceived as not available because the service he uses is not available. How can this be? The simple reason is that there are several types of outages which can impact a system or a service. One is a failure of hardware or software so that an application or service is unavailable: this is a physical outage. The other outage is a logical outage where the physical system is working but the user cannot access the service properly, that is, in the way agreed with IT at the outset. This post is based on the information in the CRC December 2014 publication High Availability IT Services by Dr. Terry Critchley. 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. |
| 4:00p | A Comparative Study: Modular Versus Legacy Containment New data center platforms are being designed and built to support large amounts of new applications. These infrastructures are created to handle powerful workloads capable of distributing data and information all over the world. In architecting the modern data center platform, administrators are striving to create an environment built on performance and efficiency. Part of the development process will always revolve around airflow and room-level control. A Mature and Proven MethodAirflow Management (AFM) is a mature and proven method for reducing IT equipment inlet temperatures, increasing cabinet, row and physical infrastructure utilization, improving cooling efficiency, and other important physical and operational parameters within the computer room space. There are various strategies and implementations for AFM, many of which are influenced by the existing conditions within the data center. Common AFM practices include, but are not limited to: sealing raised floor openings; installing blanking panels in unused cabinet u-space; sealing gaps in equipment rows; sealing the cabinet top, bottom and side rails; and installing hot aisle/cold aisle containment. Cold aisle containment, being the most simplistic and commonly implemented form of containment, is the subject of this white paper from Upsite Technologies. The effectiveness of two different types of cold aisle containment, AisleLok Modular Containment and legacy ‘full’ containment, as they relate to IT inlet temperatures and data center air conditioning efficiency, are analyzed with their corresponding results presented. Comparing Cold Aisle ContaimentAirflow management is one of the most important parameters for optimizing operations and utilization of the data center. This comparative analysis of legacy containment to the new AisleLok Modular Containment solution was actually installed on the same non-contained data center. Legacy “full” containment is a constructed enclosure encompassing selected cabinet rows or a section of the data center space. AisleLok Modular Containment is an aisle containment solution applied to individual IT cabinets and is made up of two components: Rack Top Baffles and Bi-Directional Doors. The Rack Top Baffles are applied magnetically to the top of individual IT cabinets, blocking exhaust air recirculation. The Bi-Directional Doors are applied magnetically to end-of-row cabinets, blocking the end of the aisle. Both approaches showed significant improvement to IT inlet temperatures and ability to reduce CRAH centrifugal blower speeds without exceeding ASHRAE maximum recommended 80.6° F (27° C) IT inlet temperatures. Download this white paper today to see how the results show both solutions had improved performance, but AisleLok Modular Containment shows a faster ROI and a lower total cost of ownership. |
| 4:30p | China Delays Counter-Terrorism Bill Amid Concerns from Foreign Tech Firms
This article originally appeared at The WHIR China is delaying a draft counter-terrorism bill that would require companies to keep servers and user data within the country and hand over encryption keys to the government if put into law. According to a report by Reuters on Friday, China has decided to suspend the third reading of that law, putting it on hiatus for now. White House cybersecurity coordinator Michael Daniel said that the US considers the law “as something that was bad not just for US business but for the global economy as a whole, and it was something we felt was very important to communicate very clearly to them.” Daniel was speaking about the law in a discussion at the Information Technology and Innovation Foundation last week. While it is unclear whether the bill would proceed or not, it could be picked up again at any point as only the standing committee – not the full parliament – is required to pass the law, according to an anonymous source speaking to Reuters. The initial draft of the bill was published last year, and also requires companies to censor terrorism-related content. The second draft was read last month. While the law would apply to both foreign and domestic firms, US officials believe it could unfairly target foreign companies operating in China, particularly with the new banking rules and anti-monopoly investigations in China. This story originally appeared at http://www.thewhir.com/web-hosting-news/c |
| 5:26p | VMware Data Center Comes Online in Frankfurt VMware announced launch of a data center in Frankfurt, second European location that will serve its enterprise cloud services. There are also two VMware data centers in the U.K., launched in 2013 and 2014. Germany is a strategic data center location for cloud and other IT infrastructure service providers. The E.U. in general has stricter rules about physical data location than elsewhere, and Germany is known for being one of the countries with highest concern about data sovereignty. VMware, majority-owned by the storage giant EMC, first announced plans to expand vCloud Air infrastructure to Germany last year. Amazon Web Services, VMware’s (and everyone else’s) biggest competitor in public cloud infrastructure services, announced launch of a Frankfurt data center in October last year, adding a second availability region in Europe. The other one is in Ireland. Oracle said last September it would also establish two data centers in Germany to support its cloud services. There have been anonymously sourced reports that Microsoft was planning to bring an Azure data center to the Central European nation as well. VMware’s vCloud Air services are aimed at enterprise users. The pitch is seamless integration of cloud resources in VMware-operated data centers with customers’ own VMware infrastructure hosted in their own facilities. VMware data centers that support its cloud services are usually in colocation-provider facilities. The company also has partners that provide its services hosted in their own data centers. “VMware’s new Frankfurt service location enables IT teams, developers and lines of business to take full advantage of an enterprise-class public cloud service that is local and fully compliant with local data protection regulations,” VMware CEO Pat Gelsinger said in a statement. Also this week, VMware launched into general availability vSphere 6, the latest release of its blockbuster cloud computing infrastructure software for enterprises. |
| 6:28p | Romonet Enhances Predictive Modeling With NoSQL Database Romonet, a London-based company that makes predictive modeling software for data center management, has added an element of big data analytics to its software solution that already includes its proprietary predictive modeling engine. The advantage of adding a big data architecture, with a NoSQL database at its core, is faster and more frequent predictions about the data center’s health. If there is a performance issue, the data center operator can identify and address it faster, potentially preventing customer downtime. Using software to analyze data generated by devices on the data center floor for more efficient data center management is a growing field. Last year, Google revealed its quite sophisticated approach to this problem, using a “neural network” to analyze operational data generated by its data centers and make recommendations for improvements. Typical data center monitoring and trend analysis tools act on events that have already happened, such as alarms, or historical trends, Romonet CEO Zahl Limbuwala, explained in a statement. Romonet’s software collects device data in real time, analyzes historical data, and can warn the operator of an upcoming problem. “It can actually provide a warning when device performance starts to degrade or exhibit small but important early-warning signs of failure,” Limbuwala said. Romonet’s software’s strength has traditionally been in using data about a data center to create models that provide comprehensive financial analysis of how well or poorly it performs and to predict what effects certain changes will have. The new features add real-time functionality for day-to-day data center management. The company’s engineers have also spent time to improve the way their software discerns between useful data and low-quality data generated by monitoring devices. Romonet CTO Liam Newcombe said about 35 percent of data coming out of building management systems was “junk data.” The upgraded solution cleans, verifies, and stores metered data in the NoSQL database. Data can be stored over a period of many years and used for trend analysis. |
| 7:22p | Global Capacity to Provide Direct Links to Google Cloud Global Capacity, which specializes in Connectivity-as-a-Service via an automated platform, has been approved as a Google Cloud Interconnect service provider and now allows enterprises to tap into Google’s cloud directly through its One Marketplace. Global Capacity’s One Marketplace provides real-time network pricing and ordering of private Ethernet connectivity and IP services to more than 200,000 circuits and 20,000 of its business customers. Interconnection can occur on any Google Cloud Platform service, including Google Compute Engine, Google Cloud Storage, and Google BigQuery. Google Cloud Interconnect is a relatively new program with a select list of 10 or so Carrier Interconnect providers across network and data center service providers offering direct cloud connectivity to the internet giant’s infrastructure services. Equinix, IX Reach, Megaport, International network peering play IIX, Level 3, Tata, Telx, Verizon, and Zayo are all on the list. The addition of direct cloud connectivity makes Google’s services more appealing for enterprises looking for a secure connection to public cloud to extend and complement their current setup. Amazon Web Services direct connectivity is also available through the Global Capacity marketplace. One Marketplace includes an application called Network Builder, used for designing, pricing, and ordering data connectivity between businesses and the Google Cloud Platform location. More than 9.6 million commercial addresses are available to Google Cloud Platform. “Google approved One Marketplace as part of a small group of Google Cloud Interconnect service providers, underscoring the innovation that Global Capacity and One Marketplace have brought to the data connectivity and cloud market,” said Ben Edmond, chief revenue officer of Global Capacity, in a press release. “Global Capacity will continue to transform the process of consuming network connectivity in support for Google’s continued innovation in the market.” |
| 7:47p | DataCentred Builds ARM-Powered OpenStack Public Cloud English open source cloud provider DataCentred is entering into uncharted territory, adding ARM-based servers to its OpenStack public cloud. Developed together with system level software infrastructure specialist Codethink, the cloud will run on HP M400 ARM hardware (HP’s famous Moonshot servers), but customers will have access to both Intel and ARM architectures. Processors built using architecture licensed from U.K.-based ARM Holdings power most of the world’s smartphones. But because they consume relatively little energy, there is now also a growing market for ARM servers. Numerous companies have been working on adapting the architecture for servers, but production cloud deployments are very much in their infancy. The availability of a cloud powered by ARM servers is a step forward in the development of energy-efficient clouds. Interest in ARM and other processor architectures demonstrates that there is an appetite in the market for alternatives to the domineering presence of Intel’s x86 architecture in the data center server market. Currently, there is a lot of momentum behind OpenPOWER, particularly around hyperscale and big data use cases. Born out of IBM’s POWER architecture, big service providers like Rackspace, IBM SoftLayer, and OVH are developing OpenPOWER clouds. Oracle has its SPARC chips, which can also be tied to OpenStack. One of the first service providers to have launched cloud services built on ARM servers is Online.net, the web hosting arm of French telco Iliad. Online.net designed its own servers and its own cloud architecture for the service, which came online last October. Some other examples of ARM adoption include Chinese search engine giant Baidu, which has a commercial deployment of ARM-based servers for cloud storage. Boston Limited has an ARM cloud called ARM-as-a-Service. Although there has been work and discussion about marrying OpenStack and ARM for years, DataCentred believes it is the first to market a 64-bit ARM OpenStack public cloud. The company expects to roll out the cloud into production this year. “This is probably the first example of Moonshot AArch64 running in Europe outside of HP’s development labs, and certainly the first example of generally available Moonshot backed AArch64 instances in an OpenStack public cloud anywhere in the world,” DataCentred CEO Mike Kelly said in a press release. “We know that ARM themselves are pleased to hear of this development, as a real-world deployment. OpenStack is one of the big success stories for open source software and is likely to be the environment through which enterprise migrates, in a vendor-neutral way, to take advantage of elastic cloud compute.” DataCentred and Codethink have worked together since 2014 on multi-architecture OpenStack cloud service infrastructure for customers requiring access to specific architectures to run targeted workloads. Chinese hardware giant Lenovo is developing an ARM server powered by Cavium’s 64-bit Thunder System on chip. Other major players include: Applied Micro, AMD, and Texas Instruments. |
| << Previous Day |
2015/03/17 [Calendar] |
Next Day >> |