June 15th, 2007 [by Doug Alder]
Help us to help you better. is genuinely interested in finding new ways to assist you, the customer or potential customer, to get exactly the right product for your needs. It neither serves you or RackForce well if you end up on a r that is too much or too little for your needs as either way you will end up dissatisfied. One way we are considering to achieve that is to build a questionnaire with simple yes/no questions, that you can answer online, that seek to illuminate the way you intend to use the server you are searching for. When you have completed the survey a copy will be sent to sales staff and you can then choose to either have us respond via email (we promise not to abuse your email address) or open an online chat right then with Sales and discuss it. Everyone’s time these days is important and this method will save both you and us time by getting straight to the heart of your needs. If you are not a native English speaker then this method give you the time to translate and answer the questions more easily and will greatly assist sales staff to help you further.So here’s a question for you. If we build it will you use it? Your comments and suggestions for questions that need to be asked are most welcome.
June 5th, 2007 [by Doug Alder]
[tag]Microsoft[/tag] has demonstrated a whole new “desktop” technology called [tag]Surface[/tag] that will revolutionize the way people interact with computers. Surface technology is not simply a touchscreen. Think “Minority Report” . This is the start of that kind of technology. Currently Surface is only available on a coffee table and costs about $10K but Ballmer and Gates expect to get it down to $1K in the not too distant future. What will be interesting is to see what media they can mount this system on as it requires the use of 5 near infra-red cameras to sense the screen and physical items on it such as hands, cell phones, cameras, wineglasses .Imagine putting your digital camera on or near the surface and being able to download/upload pictures with a wave of your hand. How about in a restaurant putting your wineglass down on the surface and having the restaurant’s server identify the wine, give you full information on the vintner, the varietal, the vintage and the best pairings in the menu for that wine, all accompanied by videos of the winery etc. The possibilities are endless.This is very cool technology.
June 5th, 2007 [by Doug Alder]
First let me say this post is not about whether Linux or Windows is better, both, when used properly by knowledgeable administrators, are excellent operating systems for servers.When we started selling (Virtual Private Server) plans we only offered Linux and we chose RedHat 7.3 as the OS then went to RedHat9. When RH9 reached end-of-life we switched to Fedora Core 1 (FC1) and followed that through to FC4 when we switched to . At all times we stuck with the most stable product. Stability is paramount because the amount of work required to maintain stability is always a consideration for sysadmins. Sticking with a single in offerings is also important for streamlining an operation.Control panel makers like , and always stay with the stable releases and because they need to wait until the release is stable before they upgrade their products they often appear to be be way behind current technology. Well yes they might be but again it is a matter of stability. Control panels are meant for production servers and as such you don’t put bleeding edge software on them, at least not if you want to retain your customers (or perhaps your job).If you need a server with the latest version of Apache (2.2.4 at time of writing), or the latest and greatest PHP (5.2.1) or MySQL (6.0 -Alpha) you need to get it without a control panel and install these items yourself. It will be a long time before the makers update their software to those versions and even then it will depend on the Linux distro as well.I have been asked a number of times why we dropped Fedora and went with CentOS. Well the reason is simple. Fedora has a 6 month life cycle and whether you have 10 servers or thousands of servers to maintain, trying to upgrade each one to the latest and greatest stable release every 6 months becomes Herculean task. For that reason, Fedora should not be used for production servers, it’s really a desktop Linux.
maintains any particular release of Fedora for a month after final release of the second following release. For example, Fedora 7 will be maintained until a month after Fedora 9 is released. Since new releases of Fedora occur approximately every 6 months, this translates into approximately 13 months of updates for every release. This schedule allows end users to optionally skip every other release and directly upgrade from Fedora N to Fedora N+2 release.
In short there is always a good reason for these decisions even if they are not obvious to a casual observer.
May 10th, 2007 [by Doug Alder]
What exactly is it that a infrastructure provider is promising you when they tell you that you have a , and what does it mean? If you have not already done so please read Understanding as a primer for this article. purchase from carriers typically on a 95th percentile with a certain “commit” level or as a second option they purchase bandwidth on a “capped” model. (Although it is possible to purchase on a “volume” or GB usage model this is not a standard way of buying bandwidth in a model.)RackForce offers the Carrier version of the capped rate model; for example, our dedicated 10 Mbps connection is capped at 10Mbps. This is the peak rate of the bandwidth that can be used at the server or VPS. Whether the server plugs into a 100 Mbps or a Gigabit switch port is irrelevant as the real concern is always how congested is the switch port and not what is the maximum rate or capacity of that switch port.So, that begs the question, what is congestion in this instance? The definition of congestion is based on the 95th percentile, the average sustained rate divided by the port maximum rate equals the percentage of congestion. So you could have 20, 30, 40 or more 10 Mbps servers plugged into a 100 Mbps port as long as your port congestion is within acceptable limits. RackForce believes that should be no more than 40%. That number is very important. In TCP/IP networks there is packet overhead and once you reach about 60% of capacity on a connection you start to get packet contention and from that packet loss and that leads to poor network performance.For example if there are 30 servers, each with a dedicated 10 Mbps port, all connected to a 100 Mbps switch, the lineal math calculation is 30 X 10 = 300 Mbps, but the actual average usage on that switch, expressed as a 95th percentile, may only be 30Mbps. The Carrier (data center in this case) will monitor the switch port and if the average sustained usage is 40 Mbps or higher (40% congestion) the port capacity will be increased (if the Carrier uses a 40% congestion rate policy) or will move customers to reduce the load.Check with your data center/hosting provider to see what they offer for this. If it is more than 40% consider moving your business elsewhere.In summary, how does this apply to RackForce? We employ a Carrier model of capped rate, which is a dedicated unmetered service from the switch port to the server. Beyond that, the Internet and Ethernet switching is actually a “shared” bandwidth model in general (not withstanding sophisticated architecture like MPLS, but that is not applicable in data centers); a Carrier or ISP will monitor port congestion to ensure quality of service. Most offer a 60% congestion guarantee, RackForce offers a 40% guarantee. If a port has been identified over the 40% congestion then we take appropriate action. Quality servers, quality network and quality service is what is all about.
May 9th, 2007 [by Doug Alder]
This is the way the world is going, . In our previous post we hit upon a couple of ways virtualization (Called DDS or Dynamic Dedicated Services at RackForce) can benefit server system administrators. Here are some more:
Since we handle hundreds of customer issues, changes and upgrades every week we continually see the value of this approach. If you have a DDS server you likely have already experienced its benefits and your comments would be appreciated. DDS is a RackForce strategy with the goal of delivering “On Demand” services for everything the company does. In our next blog post you will read Part I of the Roadmap so you can gain a more in depth understanding of how your Infrastructure Provider (RackForce) will give you the best opportunity for success.
April 13th, 2007 [by Doug Alder]
The
The board ignored the arguments of the International Webcasting Association and other webcasters, and apparently simply endorsed the proposal of the RIAA-associated SoundExchange royalty organization, which represents the major and some indie labels.The new rates forcewebcasters to pay for each song streamed to each user, and increase over the next few years as follows:2006: $0.0008 to stream one song to one listener2007: $.00112008: $.00142009: $.00182010: $.0019Those fees will add up quickly for larger webcasters; the Radio and Internet Newsletter (RAIN) calculates that, assuming that the average station plays 16 songs per hour, sites would have to pay “about 1.28 cents” per listener per hour using the 2006 rate, and would owe this retroactively, in addition to licensing fees going forward. RAIN’s math indicates that the rate would render Internet radio unsustainable, or at the very least, more ad-laden than terrestrial radio — and that’s before the songwriters’ licenses are taken into account:
This penalty is very injurious to anyone that is operating on an amateur basis. In all likelihood their costs will far outstrip any revenue they may generate through ads. The only real solution for such parties is to move their operations to an offshore server. If you are thinking of moving it to Canada then these two articles by Michael Geist are an absolute must read. For example he points out that SoundClick.com lists 354,675 royalty free songs using creative commons licenses. Useful to know! From the looks of it, while it won’t be free to broadcast from Canada it will work out substantially cheaper than remaining in the US.Decisions like that of the U.S. Copyright Royalty Board show a need for technology visionaries to get involved in government. Decisions like this show that the people making them are not very clued in as to the new gestalt of communications, the world, and the Internet; what can’t be done easily somewhere will be done somewhere else outside the regulatory reach of the first place. Bureaucrats, insulated from reality by their years in public service and their interactions being mainly with politicians and corporate lobbyists are out of touch with the realities of new paradigms in communications. Add to this the problem of politicians who, with few exceptions, have no concept at all of the nature of technology, and specifically the Internet (as in Senator Ted Stevens’ infamous “”The Internet is not something you just dump something on. It’s not a truck. It’s a series of tubes,” remark last June) crating legislation they don’t understand the ramifications of, and you have situations like this happening. Every country needs more tech savvy and net aware people in public decision making positions. This is integral to the future economic health of those countries.Update April 16,2007: via Kurt Hanson’s blog RAIN (Radio and Internet Newsletter) we find that:
The Copyright Royalty Board, a three judge panel responsible for the March 2nd ruling that set webcast performance royalties at their new increased rate, has denied all parties’ motion for rehearing of the ruling.The Board claimed that the motions introduced no new evidence and were therefore legally insufficient.Webcast legal vet David Oxenford, speaking at the RAIN Summit this morning, said the board “denied all the motions for rehearing on procedural grounds.”The CRB did not offer any type of clarification or additional information surrounding the minimum $500 fee per channel imposed by the new ruling.Webcasters still hold out hope for negotiations with the record industry, action in the appellate court or legislative relief.
For an idea of the impact this is going to have on small US broadcasters you should read Bill Goldsmith’s post. Bill and his wife Rebecca are the founders of Radio Paradise, a popular and pioneering Internet radio station.
April 10th, 2007 [by Doug Alder]
After these DDoS attacks, we moved FDB to bomb-proof webhosting on Rackforce, who advertise that their system is impervious to DDoS attacks. Since moving FDB to Rackforce, we’ve been hit with DDoS attacks several more times- but we at FDB would never have known about them if Rackforce Support hadn’t clued us in. Rackforce are able to divert attack traffic and move the IP of our server with just a few mouseclicks, keeping us ‘on the air’ seamlessly since we signed on.
How much is that protection worth to your online operation? $5/mo, $25/mo, $100/mo? Well at RackForce we include that protection at no extra cost on all accounts. Now that’s valueWhile we certainly appreciate the compliment from our customer I would like to point out that no data center can ever claim to be “
RackForce is committed to giving its customers the best service for the price and we are continuously upgrading our network and our offerings to meet that commitment. Stay tuned.
April 3rd, 2007 [by Doug Alder]
While using green hydroelectric power(see previous post) is important, reducing the amount of power used is also important for the environment. Doing so frees up power for other customers which reduces the amount of non-green power that would otherwise need to be purchased on the spot market to meet those demands. Done in sufficient quantities it frees up green power for export to the spot market. now uses new quad-core server technology from IBM to give its customers more processing power and reliability while at the same time reducing its electricity needs due to the new efficiencies IBM has built into this technology.The new servers are three times faster than the servers Rackforce had been previously using, while using 30 percent less electricity. Rackforce has purchased 320 IBM System x servers in 2006 for $1.2 million, and expects to purchase 600 more this year.“One of our highest costs is powering our data centres, and anything we can do to save electricity makes a big impact on costs,” said Tim Dufour, President of Rackforce. “Our first priority is to provide our customers with the best value and the most reliable hosting available, and the second is being as environmentally responsible and energy efficient as possible. Our new IBM quad-core servers help us achieve those priorities.”
March 30th, 2007 [by Doug Alder]
March 22nd, 2007 [by Doug Alder]
Notice to all RackForce customers who have their domains registered through . has terminated RegisterFly’s effective March 31, 2007. Copied below is the formal announcement located here on the ICANN site.
ICANN today issued a formal notice of termination of RegisterFly.com’s Registration Accreditation Agreement (RAA).ICANN has issued a letter to RegisterFly [PDF, 902K] indicating that it will cease operating as an ICANN-Accredited Registrar on March 31, 2007. Under the terms of the Registrar Accreditation Agreement (RAA), ICANN must provide 15 days written notice to RegisterFly of its intention to terminate.Effective immediately ICANN has terminated RegisterFly’s right to use the ICANN Accredited Registrar logo on its website.Between now and 31 March RegisterFly is required to unlock and provide all necessary Authinfo codes to allow domain name transfers to occur. Any and all registrants wishing to transfer away from RegisterFly during this period should be allowed to do so efficiently and expeditiously.”Terminating accreditation is the strongest measure ICANN is able to take against RegisterFly under its powers,” Dr. Paul Twomey, President and CEO of ICANN said today.”ICANN has been frustrated and distressed by recent management confusion inside RegisterFly,” Dr. Twomey, President said. “I completely understand the greater frustration and enormous difficulty that this has created for registrants.”When the Agreement is terminated, ICANN can approve a bulk transfer of all current RegisterFly domain names to another ICANN accredited Registrar.”Of course, RegisterFly does not have to wait till then. They can request ICANN to approve a bulk transfer immediately. I call on RegisterFly to act in the interests of registrants and seek such a transfer from us straight away,” Dr. Twomey said.ICANN intends to hold a forum to discuss the reform of the Accreditation policy and process at its Lisbon meeting in a week’s time*.A set of questions and points to inform the discussion will be made public prior to the Lisbon meeting.* The Lisbon meeting is one of three meetings held a year by ICANN to meet with global stakeholders. It will take place from 26-30 March 2007.Media Contacts:Jason KeenanMedia AdviserICANN ( USA)Ph: +1 310 818 9072E: jason.keenan@icann.orgInternational: Andrew RobertsonEdelman ( London)Ph: +44 7921 588 770E: andrew.robertson@edelman.com
| « Older Entries | Newer Entries » |
