Technology News You Can Use

May 2007 – Issue XXVI

In This Newsletter

• Letter from the Chair
• What is Colocation and Why Should I Care? Part III
• Open Source Alternatives: Linux on the Desktop
• Great Googl[e]y Moogly  - Part II
• Get LAN-Like Performance from Your WAN
• Contact Us

Letter from the Chair, Zandra Pulis, CPS Energy

I was thinking about our May Tech meeting and about all the exciting things going on with the Tech Council. I believe that the Tech Lunch Bunch is one of the most exciting events on our horizon because it provides an outreach to the North Chamber members - especially small businesses. These members have an opportunity, by virtue of their North Chamber membership, that they may not otherwise have - the opportunity to learn about some broad IT issues that affect their business. I sincerely commend the Tech Committee for developing this great program.

The first Tech Lunch Bunch is schedule for June 13th at the North Chamber. Cost is $10 and includes a boxed lunch. What a deal! You can click here to learn more. Please spread the word about this event to your North Chamber friends, especially those well suited to benefit from the program. You can register online or call the North Chamber at 344.4848.

Have a great week and GO SPURS GO!

Greetings

Zandra Pulis

CPS Energy

Top

What is Colocation and Why Should I Care? Part III
Paul O'Bannon, Nugget Enterprises

Let’s revisit for the last time our scenario from the first article:

You’re a small business.  All of your customer data is on the server at the office.  The air conditioner failed during the night, and the heat in your “computer closet” has overtaken the gear and it fails.  Your data is gone.  Will your business survive?  Can the data be recovered?  Could I have prevented the panic somehow?  To a business that uses a colocation facility, the answer to all these questions is a resounding “Yes.”

Now, let’s think about some other scenarios.  Remember our ice storm back in January?  How many businesses lost productivity or perhaps simply “closed shop” because staff couldn’t get to the office?

Business continuity is a subject that runs parallel with disaster recovery.  Business continuity is that portion of the plan that allows a business to continue to run at as close to 100% as possible.  The things that must be taken into consideration are:

1.      How will the phones get answered?

2.      How can I get access to my business systems?

3.      How can I get access to my business data?

I’m not going to pretend that anyone can base a complete business continuity plan on this article.  It’s only designed to help you think along these lines, and how colocation fits in.  Many businesses have made these evaluations and have placed computer gear in colocation facilities for this exact reason:  Keeping the business running under abnormal circumstances.  By securing their gear in a colocation facility, they are quite assured that their equipment is up and running, and that their data is current (assuming a quality backup plan).  And because the gear is there, staff can work from home, or perhaps a temporary office arrangement.  All they need is an internet connection, and they are back in business.

To wrap this whole thing up:  Who needs to use a colocation facility?  Any business that relies heavily on their business data.  Any business that needs to stay in business when things go wrong.  Any business that understands that loss of business data can mean loss of the business.  Any internet-based business.  Any business.  Period.

Top

Top

Open Source Alternatives: Linux on the Desktop, Steve Weingart, HighTek Innovations

In late 2004, a memo from CIO Bob Greenberg said that IBM Chairman Sam Palmisano has "challenged the IT organization, and indeed all of IBM, to move to a Linux-based desktop before the end of 2005."  IBM has not yet completed a full changeover to a Linux-based desktop, but IBM has demonstrated its commitment to Linux on many fronts.  Earlier this year PSA Peugeot Citroën, the second-largest automobile manufacturer in Europe , and Novell signed a multi-year contract allowing the deployment of up to 20,000 Linux desktops plus 2,500 Linux servers from Novell.

Why would some of the world’s largest and best financed companies do this?  Here are a few of their reasons: “We found SUSE Linux Enterprise Desktop to be well supported and extremely user friendly,” said an IT representative for PSA Peugeot Citroën.  The Citroën representative went on to say, "In addition, SUSE Linux Enterprise Desktop integrates seamlessly in our Windows-based infrastructure."  Ron Hovsepian, Novell's president and CEO, said, "PSA Peugeot Citroën has discovered the value a market-ready Linux platform can bring to their business, … and help the organization reduce costs while advancing its next-generation IT architecture."

So, how does this apply to your small business?  Can a Linux desktop actually reduce your IT costs?  Will user productivity remain the same, or better yet, improve?  IBM, Peugeot Citroën and others (including national governments and school districts all over the world) believe that the answers are ‘YES.’

From a cost standpoint the biggest difference between Open Source and Traditional business models are licensing charges.  Traditional software business models rely on charging for licenses and upgrades, then charging more for support.  The Open Source business model does away with the license charges and charges only for support.  You ask, “How can that possibly work”?  It does.  SUSE Linux is owned by Novell, a giant in the computer industry.  IBM has been shipping Linux software loaded on servers and desktops for several years. Dell has been shipping Linux software on servers and is about to start shipping Linux on desktops and laptops for consumers (it has already been available for business), and Open/Star Office, one of the jewels in the crown of Open Source, is owned by Sun Microsystems, another giant.

In practice, support for Open Source operating systems and applications is available in at least two ways.  You can pay the manufacturer for support, by the incident, or under contract, just as it’s done with Traditional software.  You can also pay nothing to the manufacturer for the software and self support (if you have the expertise), or you can contract for support with a service company.  All of these methods work.  By going this way you will typically save between $100 and $400 on the operating system, and that much again for your office suite, per desktop by going with a Linux desktop.

Training is another concern, “How are my users going to deal with this new environment”?  In practice it turns out that the modern Linux desktops are sufficiently familiar looking that little or no training is needed for usual office functions such as document preparation/word processing, web browsing and using applications.  Since more and more applications are becoming web browser based, users will not have to contend with many unfamiliar interfaces.  But, like support, training is available (again from the manufacturer or a service company); and it’s not likely that the amount of training needed will be any more than if you upgrade to Windows Vista.

Another area where you can save significant costs is the hardware.  Linux uses fewer resources.  An ‘office class’ PC will exhibit great performance with any one of the popular Linux desktop versions.  If you are planning to upgrade to Windows Vista, consider the recommendations for the base hardware, and you will see that you will need a much more powerful PC to do regular work, and you will need even more if you want to take advantage of all of the new features.

Lastly, you will gain freedom from lock-in.  You will no longer be tied to one vendor and that vendor’s pricing schedule for licensing and support.  You will have the freedom to make choices that can save you money while keeping up productivity.

Look into Linux, you just might like it.

Top

Great Googl[e]y Moogly  (Part II), Johnny Harvey

In part two of this two part series, we’ll explore the many facets you need to understand to get your website into a respectable range of search results when people search for your website on Google, Yahoo or any one of hundreds of other search engines available on the Internet.  To view last month’s Part 1 article, visit http://www.northsachamber.com/etech_news.php and click  the April 2007 E-news article link.

Page Structure and Content

Page structure directly influences the understanding of content by humans and indexers.  Whether writing a document for print or for the Web, the presentation should be hierarchical.  That is, each page should have a title, headers, sub-headers, italics, and similar formatting.  This topography provides the structure and meaning mortals need to read and understand what a writer is attempting to convey.  Beneath the veneer, it enables spiders and bots to do the same, but by means of XHTML markup rather than visual cues.  This is possible because of the semantics tags provide to these programs.  The title tells us, for example, about a page’s topic.  We know this because the font is larger, bolder, and often a type differing from the body text.  When wrapped in an <h1> header tag, search engines have the same understanding, placing a higher importance on it above content enclosed in tags used for sub-headings and paragraphs, prioritizing each accordingly for the algorithms used to determine rankings on SERPs.

The placement of content also has meaning.  For us, we depend on the visual layout either provided by tables or cascading style sheets (CSS), of which the latter is highly preferable.  Without styling, the content would display unformatted and largely unintelligible.  For software agents, semantics comes from the unformatted stream of text pooled with structural tags.  In both cases, there is one facet, regardless of whom or what is reading it, necessary for understanding – the content must be in a certain order.

Visually, we are concerned with the physical arrangement on the page.  We arrange text into paragraphs and sections in accordance with current best practices and adorn the page with imagery reflective of our tastes.  Spiders and bots couldn’t care less about the visual acuity; software agents consume text the same way we add it to a page: left to right, top to bottom.  The primary reason for ensuring content is ordered in a certain manner for agents is because the higher on the page, the closer to the <body> tag, the greater importance attributed to keywords and surrounding text.  As an aside, if you choose not to use the description meta tag in your pages, as previously discussed, what indexers read in first is of particular importance.

There are two ways to have content preferential arranged for humans and software agents: (1) build the page in that order or (2) use CSS to arrange it.  Truly optimizing your pages for indexing renders the first option unrealistic if you intend for your site to have any value for your visitors.  Using CSS, you are able to construct the markup so that it’s order independent.  This means we can present a right-hand column on the left and a left-hand column on the right, while underneath leaving it unchanged to the benefit of the spiders and bots.  Let’s try an example.

Imagine having a simple two-column layout.  Looking at the page in a browser, you see In the left column a mash up of links, comments, and quotes.  In the right column, you see the main content, a review about the latest movie releases.  On its face, it seems software agents would consume the content beginning with the mash up in the left column.  However, because we formatted the page using CSS, the right column is actually the left column when viewed at the code level; it’s only occupying the space on the right and vice versa for the other column.  Because the main content is in fact within the left column, the closest to the top-left corner of the page, spiders and bots read in the review first and, therefore, attribute higher scores to the keywords located in that section.

SEO is an iceberg of a topic and we’re still above the water line.  Google SEO to discover the plethora of information available online.

Top

Get LAN-Like Performance from Your WAN, Brett Rushton; Calence

Applications on the WAN: Slow and Getting Slower

Wrestling with application performance over the Wide Area Network (WAN) is a common problem, particularly in today’s IP networks with distributed work forces, centralized data centers and inefficient network applications.   For many users, slow application performance is more than a petty annoyance; it reduces productivity, erodes revenue and frustrates employees.

Looking forward, this problem is going to get worse before it gets better.  We live and work in a global economy where users are more dispersed, systems and applications are being centralized, and  WAN links are more congested with data-intensive applications.

Underlying Technical Issues

The biggest hurdles to providing WAN users with corporate LAN-like performance are latency, bandwidth and TCP/IP.  Simply stated, latency is the time it takes for data to travel across the WAN.  This delay is a function of distance – the longer the distance, the higher the latency.

Even though latency is only a fraction of a second, it becomes a sizable bottleneck if your applications employ “chatty”, inefficient code with many requests and responses strung together.  Each request-response pair is hit with that latency delay, which means applications that issue hundreds of them per transaction, which is not uncommon, really slow down.  This effect is one of the reasons why users on the corporate LAN, typically much closer to the corporate data center and application servers, don’t seem to complain as much about slow applications.

Bandwidth becomes the issue when applications send large amounts of data, such as high-resolution graphics, that exceed the capacity of the WAN connection.  Often, applications are developed in high-speed LAN environments where bandwidth is essentially unlimited.  Code is not optimized, or tested, to be efficient in lower speed WAN environments.

Compounding this problem is the operations of TCP/IP.  TCP/IP - developed years ago for low-bandwidth, unreliable, networks - prevents the optimal use of today’s networks.  Its flow-control and congestion-avoidance mechanisms, limits throughput well below the physical circuit speed.

Generally, a CIO’s options for improving application performance are limited to:

Rewriting the application, which is often cost prohibitive

• Increasing WAN bandwidth, which is expensive and may not solve the issue
• Deploying next generation WAN acceleration and optimization solutions

Application Acceleration and WAN Optimization

Over the last 12-18 months, network-based solutions for “WAN optimization” and “application acceleration” have matured to become a viable cure for these WAN performance woes – a nice alternative to rewriting or re-architecting your applications.  This market grew to $500 million in 2006 and is expected to continue to grow to over $2 billion by 2009, according to estimates from Cisco, Gartner and IDC. 

There are a number of different vendor solutions available on the market.  Each has its strengths and weaknesses.  As you consider what WAN optimization/application acceleration solution to deploy in your environment, there are three things to consider:

1. How will you measure the actual performance improvement?  This requires you to understand your application architecture, traffic flows and protocols and to baseline your current response times.
2. How do you make sure you handle all of your user traffic?  This requires you to appropriately size the solution architecture according to metrics, such as WAN bandwidth, number of users, number of TCP connections and to consider needs for future scalability
3. How and where will this integrate into your network environment?  Network design elements include routing, monitoring and security all need to be taken into account

The Solution – Cisco Wide Area Application Services (WAAS)

Cisco’s WAAS solution is a comprehensive solution that addresses each of the underlying technical issues mentioned earlier, to provide application acceleration and bandwidth optimization for any TCP-based applications on your WAN.  This solution allows you to:

§         Deploy centralized applications with confidence

§         Consolidate costly branch servers and storage into data centers

§         Optimize the use of WAN bandwidth

§         Facilitate centralized data backups

Most importantly, WAAS integrates transparently with the network, preserving TCP information to maintain functions such as security, QoS, visibility and end-to-end monitoring.

Cisco WAAS achieves superior results by leveraging a hardware footprint in the remote office and in the data center, overcoming application performance problems in WAN environments. It optimizes TCP flow, reduces overall WAN consumption based on data redundancy and compresses all outbound traffic. WAAS also optimizes specific applications, which eliminates unnecessary chatter and saves additional WAN bandwidth.  It allows the pre-population of WAAS edge devices, putting popular content closer to branch users, and enables access to copies of shared files in the event of a WAN outage.  .

Testing by Cisco and by third parties has shown that WAAS improves performance for virtually any TCP-based application, through its TCP flow optimizations.  Many applications enjoy the benefits of WAAS caching and compression, while still others, like file-sharing applications, reap a large performance boost from its application-specific optimizations.