Helping a colleague this week, we uncovered some odd behavior with a site whose performance he was analyzing. Upon first glance, it was clear that this site had a performance issue – they had HTTP persistence disabled. Immediate red flag in the areas of network overhead and geographic latency.

Further digging exposed something more sinister. It seems that HTTP persistence was only disabled for browsers with MSIE in the user-agent string. Even if the user-agent string was just MSIE, HTTP persistence was off.

The customer was very forthcoming and sent us their standard httpd.conf file. This showed no sign of the standard (and frustrating) global disabling of persistence for Internet Explorer.

Finally, it came to us. The customer had provided a simple network diagram, and there, just before packets hit the Internet, was a Layer 7 firewall. How did we know the Layer 7 firewall was the likely cause? Because this device was also the one that provided compression for the content going out to customers.

A Layer 7 firewall happily rewrites HTTP headers to reflect the nature of the compressed content (content-length or transfer-encoding: chunked) and to add the gzip flag (accept-encoding:gzip). Since this device was already doing this, it was pretty clear to us that it also had a rule that disabled HTTP persistence for anything with MSIE in the user-agent string.

This was a fine example of the complexity of the modern Web application infrastructure. In effect, there were two groups with different ideas of how Internet Explorer should be handled at the network layer, and neither of them seems to have talked to the other.

When you have a Web performance problem, indulge in a thought experiment. Create an imaginary incoming Web request and try to see if you can follow it through all the systems it touches on your system. Put it on a whiteboard, a mindmap, whatever works.

Then invite the system architects and network engineers in and get them to fill in the gaps.

No doubt that will lead to the “ah ha!” moment. If nothing else, it’s a good excuse to put pizza on the company card. But I have no doubt that you will walk away with a better understanding of your systems, which will make it easier for you to talk to all the people responsible for keeping your systems running.

TAKEAWAY: Just because the part of the Web application you work on is working fine, it may be affected by other components that are not tuned or configured for performance. Get to know the entire application at a high level.

Every so often, you wake up and realize that the world has changed around you. Or, to say it better, your view of the world has changed so profoundly, but also so subtly and slowly that it is imperceptible unless you take the time to look back at where you came from.

Six years ago, if you had asked me what the most important problems in Web performance were, I would have reeled off a list that was focused on technology and configuration: HTTP compression, HTTP persistent connections, caching, etc. In fact, six years on, these are still the concepts that dominate Web performance conversations.

Slowly, glacially, shaped by six years of working with customers and clients, listening to the Web performance conversations that flow across the Web and within companies, I realize that technology is only one component of the Web performance solution.

Web Performance is NOT Just Technology

Most organizations focus too much of their efforts on solving the technical problems because they are discrete, easy to track, and produce quantifiable results.

Fair enough.

But a highly tuned engine with a rusted chassis, four flat wheels, and a voided warranty still has a Web performance problem, even if it is technically sound.

The complexity of the issue arises from the terminology used. Web performance, in current parlance, refers almost completely to the delivery of the site in an appropriate and measurable manner.

Web performance is not simply the generation and delivery of HTML and other objects. Web performance is conversation that defines the basic nature of any Web site.

Approaching Web performance, as I had for so many years, as a technical problem with a discrete solution overlooks the true nature of Web performance. A culture of effective Web performance absorbs a number of different inputs, and then ensures that the site performs across many different vectors, not just the two-dimensional response/success over time graph.

Web Performance is Culture and Communication

Web performance is an issue of culture. And at the root of all cultures lies communication.

The Web performance conversation has three components, each one shaping the potential response to the problem and providing elements of the solution.

1. Technical Capabilities

Technical organizations spend a great deal of their time defining what they can’t do. In an organization that has a culture of effective Web performance, the technical teams provide clear definitions of the current capabilities, and clearly demonstrate how far they can take the organization down the chosen path, hopefully without spending all of the company’s treasure.

2. Business Objectives

Just as the technical organization has to define what they can do with what they have, the business organization has to come to the table with a clear definition of what they want to achieve. If a business goal is clearly stated to the technical team, then a conversation about where there may be challenges and opportunities can occur. When business and IT talk and listen, a company is becoming far more effective at delivering the best site they can.

3. Customer Expectations

Neglected, forgotten, nay, even ignored, the role of the customers’ expectations in the Web performance equation is just as critical as the other two participants. With clear business objectives and defined technical capabilities, a site can still be seen as a Web performance failure if the expectations of the customer are not met. And it is not simply listening to customer and providing everything they want. It’s understanding why they need a feature/function/option in order to be more successful at what they do, and balancing that against the other two players in the conversation.

Now What?

But where does an organization that wants to take Web performance beyond the technical problem, and into the realm of the strategic solution go?

Do a search on any search engine and you will find page upon page of technical solutions to a supposedly technical problem. Web performance is not solely a technical problem. In many cases, the site is configured and tweaked and tuned and accelerated to such a degree that you have to wonder if is under-performing out of spite more than any other reason.

Scratch the surface. Look beyond the shiny toys and massively-scaled infrastructure and you will find that technology is not the issue. The demand placed on the site by the business are bogging the site down in ways that no amount of tuning could improve.

Perhaps the business goals of the site, the need to support the business, have pushed the technology to its breaking point or beyond, but the technology team cannot clearly articulate what the problem or solution is.

Maybe customers, used to competitors delivering one level of Web performance and experience are simply not happy with the site, no matter how tuned it is and how clearly the call to action may be.

Making a Web site perform effectively means stepping back and asking some key questions:

• Why do we have a site?
• How does this site help our business?
• Why do our customers use our site?
• Do we like using our site?
• What are our competitors doing?
• What are the best Web companies doing?

These seem like silly questions. But you may be surprised by the differing answers you get.

And from there, the conversation can start.

Takeaways

Simply put, Web performance is not about understanding how to make your site faster. Web performance is about understanding what you can do to make your site better. An effective Web site is one that is shaped by a culture of effective Web performance.

Striving to make a better, more effective Web site may lead to such profound cultural and organizational changes that the process ends up making a better company. A company where the Web site is seen as an active conversation shared with employees, shareholders, investors, and customers.

A conversation where you explain what can be done, why you are doing it, and how you will do it. A conversation where you listen to what must be done, how it is expected to work, and what the customer defines as success.

So when you wake up six years from now, and realize that the day you stopped treating your Web site as a technical problem that needed to be fixed, and started seeing it as an opportunity to create a more effective business, I hope you smile.

About four years ago, I published/posted a number of indices related to GrabPERF data. This is a re-launch of that, starting with the GrabPERF Search Performance Index.

The methodology of the Search Performance Index is straightforward: A number of key search providers are selected and HTTP GET requests are sent that directly pull the results page of a query page that is searching for ‘New York’.

This is a simple process and one that is a close approximation of the actual search process that hundreds of millions of people perform every day.

alias                       Response Time  Success  Attempts  Success Rate
--------------------------  -------------  -------  --------  ------------
SEARCH - Google Blogsearch      0.2822545    27242     27244       99.9927
SEARCH - Google                 0.3151932    27228     27247       99.9303
SEARCH - Live (Microsoft)       0.3840097    27245     27246       99.9963
SEARCH - Indeed                 0.4112960    27240     27241       99.9963
SEARCH - Yahoo                  0.4574381    24175     24175      100.0000
SEARCH - Altavista              0.4592764    23922     23922      100.0000
SEARCH - Cuil                   0.6757475    23963     23967       99.9833
SEARCH - AOL                    0.7822945    23913     23913      100.0000
SEARCH - Ask                    0.9025220    24157     24163       99.9752
SEARCH - Technorati             0.9053472    27219     27234       99.9449
SEARCH - Amazon                 1.3251402    27245     27251       99.9780
SEARCH - Baidu                  1.7409345    23777     23799       99.9076
SEARCH - Blogdigger             1.8960633    25106     26354       95.2645
SEARCH - BlogLines              2.0238809    27233     27248       99.9450
SEARCH - IceRocket              2.1233684    24147     24177       99.8759
SEARCH - Blogpulse              2.4144131    27019     27247       99.1632

As can be seen in the data, there is a substantial degree of difference in the response time results, but the big three search providers (Google, Yahoo, and Microsoft Live) were in the top five. As well, all but three of the providers (Blogdigger, IceRocket and Blogpulse) had success rates (availability) of 99.9% or higher.

I don’t expect to see much in the way of change in these results, but I will post them for comparison each week.

Today I would like to announce the availability of the GrabPERF Search Performance Index.

The goal of the index is to provide performance metrics for a group of search providers around the world. The results are based on a direct HTTP GET request being made for the search results page by the GrabPERF Agent.

Currently only live data is available. In the near future, archival results on a week by week basis will be made available.

If there is a search provider that has been missed, please contact the GrabPERF team.

Why GrabPERF?

About four years ago, I had a bright idea that I would like to learn more about how to build and scale a small Web performance measurement platform. I’ve worked in the Web performance industry for nearly a decade now, and this was an experimental platform for me to examine and encounter many of the challenges that I see on a daily basis.

The effort was so successful and garnered enough attention during the initial blogging boom that I was able to sell the whole platform for a tiny (that is not a typo) sum to Technorati.

The name is taken from another experimental tool I wrote called GrabIT2 which uses the PHP cURL libraries to capture timings and HTML data for HTTP requests. It is an extension of my articles and writings on Web performance that started at Webperformance.org, and that have since moved to this blog.

What is GrabERF?

GrabPERF is a multi-location measurement platform, based on PERL, cURL, PHP, and MySQL that is designed to

• Measure the base HTML or a single-object target using HTTP or HTTPS
• Report the data to a central database (located in the San Francisco Area)
• Report the data using a GUI or through text based download

Why not Full Pages with all Objects?

Reason 1: I work for a company that already does that. Lawyers and MBAs among you, do the math.

Reason 2: I am an analyst, not a programmer. The best I can say about my measurement script is hack job.

Why is the GrabPERF interface so clunky?

See reason 2 above.

If you want to write your own interface to the data, let me know.

Why has the interface not changed in nearly three years?

The current interface works. It’s simple, clean, and delivers the data that I and the regular users need to analyze performance issues. If there is something more that you would like to see, let me know!

I like what I see. How can I host a measurement location?

Just contact me, and I can provide you with a list of PERL modules you will need to install on your linux server. In return, I need a static IP address of the machine hosting the measurement agent.

How stable is GrabPERF?

Most of the time, I forget it’s even running. I have logged onto the servers and typed in uptime and discovered that it’s been 6 months or more since the servers have been re-booted.

It was designed to be simple, because that’s all I know how to do. The lack of complexity makes it effectively self-managing.

Shouldn’t all systems be that way?

What if my question isn’t asked / answered here?

Your should know the answer to this by now: contact me.

Introduction to the Series

This is the first of a four-part series focusing on the reasons why companies measure their Web performance. This perspective is substantially different than ones posited by others in the field as they focus on the meat and potatoes reasons, rather than the sometimes more difficult to imagine future effects that measurement will bring.

Reason One: Customer Generation

It is critical that a company be able to show that their Web services are superior to others, especially in the third-party services and delivery sectors of the Web. In this area, Web performance measurement is key to demonstrating the value and advantage of a service versus the option of self-delivering or using another competitor’s service.

Comparative benchmarking that clearly demonstrates the performance of each of the competitive services in the geographic regions that are of greatest interest to the prospect is key to these Web performance measurements. To achieve truly competitive benchmarks and prove the value of a service, measurements must be realistic and flexible.

In the CDN field, a one object fits all approach is no longer valid. CDNs are responsible for delivering not just images or static objects, but may also host an entire application on their edge servers, serving both HTTP and HTTPS content. In other cases, the application may not be hosted at the edge, but the edge server may act a s a proxy for the application, using advancing routing algorithms to deliver the visitor the requested dynamic content more quickly (in theory) than the origin server alone.

This complex range of services means that a CDN has to be willing to demonstrate effective and efficient service delivery before the sale is complete. A CDN has to be willing to expose their system not just to the backbone-based measurements offered in a traditional customer generation process, but to take measurements from the real-user perspective.

Ad-providers have to be willing to show that their service does not affect the overall performance of the site they are trying to place their content on. Web analytics firms face the same challenge. Web analytics firms have one advantage: if their object doesn’t load properly, it may not effect the visitor experience. However, neither ad-providers nor Web-analytics providers can hide frow Web measurement collection methods that show all of the bling and the blemishes.

Using Web performance measurements to generate customers is a way that a firm can clearly show that they have faith enough in their service to openly compare it to other providers and to the status quo.

But woe be the firm who uses Web performance metrics in a way that tries to show only their good side. Prospects become former prospects very quickly if a firm using Web performance data to generate new business is found to be gaming the system to their advantage. And it will happen.

Customer Generation is a key method that Web performance measurements are used by firms to clearly show how their service is superior to what a prospect currently has, or is also considering. However, this method does come with substantial caveats, including

• The need to measure what is relevant
• The need to measure from where the prospect has the greatest interest
• The need to consider that gaming the system to show advantage will cost a firm in the end.

The GrabPERF Black Friday Dashboard is done for another year and there were two performance victims that suffered the most at the hands of the onslaught of bargain-hunters in the area of Web performance.

Some caveats that I need to mention about the GrabPERF measurement methodology.

1. Only the base HTML file of each site is measured.
2. Only the base HTML of the homepage is measured. This means that any issues that arose in the shopping process were not captured.

All of the sites in the GrabPERF Holiday Retail Measurement Index can be continually monitored on the GrabPERF Black Friday Dashboard. This page will be available until January 1 2009.

That said, the two primary performance victims this year are HP Shopping and Sears. We focus here on those that did not do that well because sites who have met the Web performance challenge and survived to fight another year are not as interesting from a learning perspective.

HP Shopping

HP Suffered the greatest response time problems, by effectively becoming unresponsive as of 09:00 EST. The greates affect on overall response time came as a result of the First Byte time metric which is a solid proxy for measuring the server or application load, as it is the time between the initial client HTTP request and the server’s HTTP response.

Factored into the poor performance analysis is the fact that GrabPERF only captures data for the base HTML object. If the performance seen here is carried over to the download of all of the graphical content on the page, I would be surprised if anyone was able to make any kind of purchases on the HP web site on Black Friday.

Today, performance has returned to substantially lower levels, indicating that this application was simply not ready for the amount of traffic it received, or ran into a completely unexpected issue when the load increased.

Recommendation for 2009: Load Test the application using this year’s traffic metrics as a baseline for validating the scalability of the application.

Sears

Sears is a returning visitor from last year’s Black Friday measurements. Unfortunately, they return for exactly the same reason that they were on last year – scaling/capacity issues that appear as errors.

And these are the worst kind of errors. As can be seen in the graphic below, the Sears Web site announced to the whole world that they had over-reached and that they could not handle the incoming volume of traffic.

What is interesting is that Sears owns properties that survived the day very well, namely Lands End. The question that must be posed is why does the parent site fail so badly when the child sites handle the traffic without difficulty?

Recommendation for 2009: Load testing for capacity, and meeting with the Lands End team to understand what they are doing to handle the load.

When starting with new clients, finding the low-hanging fruit of Web performance is often the simplest thing that can be done. By recommending a few simple configuration changes, these early stage clients can often reap substantial Web performance improvement gains.

The harder problem is that it is hard for organizations to build on these early wins and create an ongoing culture of Web performance improvement. Stripping away the simple fixes often exposes deeper, more base problems that may not have anything to do with technology. In some cases, there is no Web performance improvement process simply because of the pressure and resource constraints that are faced.

In other cases, a deeper, more profound distrust between the IT and Business sides of the organization leads to a culture of conflict, a culture where it is almost impossible to help a company evolve and develop more advanced ways of examining the Web performance improvement process.

I have written on how Business and IT appear, on the surface, to be a mutually exclusive dichotomy in my review of Andy King’s Website Optimization. But this dichotomy only exists in those organizations where conflict between business and technology goals dominate the conversation. In an organization with more advanced Web performance improvement processes, there is a shared belief that all business units share the same goal.

So how can a company without a culture of Web performance improvement develop one?

What can an organization crushed between limited resources and demanding clients do to make sure that every aspect of their Web presence performs in an optimal way?

How can an organization where the lack of transparency and the open distrust between groups evolve to adopt an open and mutually agreed upon performance improvement process?

Experience has shown me that a strong culture of Web performance improvement is built on three pillars: Targets, Measurements, and Involvement.

Targets

Setting a Web performance improvement target is the easiest part of the process to implement. it is almost ironic that it is also the part of the process that is the most often ignored.

Any Web performance improvement process must start with a target. It is the target that defines the success of the initiative at the end of all of the effort and work.

If a Web performance improvement process does not have a target, then the process should be immediately halted. Without a target, there is no way to gauge how effective the project has been, and there is no way to measure success.

Measurements

Key to achieving any target is the ability to measure the success in achieving the target. However, before success can be measured, how to measure success must be determined. There must be clear definitions on what will be measured, how, from where, and why the measurement is important.

Defining how success will be measured ensures transparency throughout the improvement process. Allowing anyone who is involved or interested in the process to see the progress being made makes it easier to get people excited and involved in the performance improvement process.

Involvement

This is the component of the Web performance improvement process that companies have the greatest difficulty with. One of the great themes that defines the Web performance industry is the openly hostile relationships between IT and Business that exist within so many organizations. The desire to develop and ingrain a culture of Web performance improvement is lost in the turf battles between IT and Business.

If this energy could be channeled into proactive activity, the Web performance improvement process would be seen as beneficial to both IT and Business. But what this means is that there must be greater openness to involve the two parts of the organization in any Web performance improvement initiative.

Involving as many people as is relevant requires that all parts of the organization agree on how improvement will be measured, and what defines a successful Web performance improvement initiative.

Summary

Targets, Measurements, and Involvement are critical to Web performance initiatives. The highly technical nature of a Web site and the complexities of the business that this technology supports should push companies to find the simplest performance improvement process that they can. What most often occurs, however, is that these three simple process management ideas are quickly overwhelmed by time pressures, client demands, resource constraints, and internecine corporate warfare.

There is clear dissatisfaction with the current state of marketing among the social media mavens.

• Fred Wilson and Union Square Ventures are looking for companies to invest in to take advantage of this.
• BuzzLogic releases their conversational ad service.
• The Inquisitr moves from AdSense to Technorati Media, indicating a potential shift at b5 Media.
• Lookery is providing demographic information that is relevant to social media providers and shaping ad presentation that way
• A completely off-the-radar person such as myself realizes that there is gold in what holds the attention, and that’s not always the latest and greatest

So what can be done? Jeff Jarvis points out that the problem lies with measurement. I agree, as there is only value in a system where all of the people involved agree on what the metric of record will be, and how it can be validly captured.

Currently CPM is the agreed upon metric. In a feed based online world, how does a CPM model work? And, most importantly, why would I continue to place your ads on my site if all your doing is advertising to people based on the words on the page, rather than who is looking at the page and how often that page is looked at.

In effect, advertisers should be the ones thrying to figure out how to get into the community, get into the conversation. As an advertiser, don’t you want to be where the action is? But how do you find an engaged audience in an online world that makes a sand castle on the beach in a hurricane look stable?

The challenge for advertisers is to be able to find the active communities and conversations effectively. The challenge for content creators and communities is to understand the value of their conversations, the interactions that people who visit the site have with the content.

In effect, a social media advertising model turns the current model on its head. Site owners and community creators gain the benefit of being attractive to advertisers because of the community, not because of the content. And site owners who understand who visits their site, what content most engages them, how they interact with the system will be able to reap the greatest rewards by selling their community as a marketable entity.

And Steven Hodson rounds out the week’s think on communities by throwing out the subversive idea that communities are not always free (as in ‘beer’, not as in ‘land of’). If a community has paid for the privilege of coming together to participate in communal events and discussions, then can’t that become an area for site owners to further control the cost of advertising on their site?

While the benefit of reduced or no marketing content is the benefit of many for-pay communities, this benefit can be used by site owners by saying that an advertiser can have access to the for-pay community at the cost of higher ad rates and smaller ads. The free community is a completely different set of rules, but there are also areas in the free community that are of higher value than others.

In summary, the current model is broken. But there is no way to measure the value of a Twitter stream, a FriendFeed conversation, a Disqus thread, or a Digg rampage. And until there is, we are stuck with an ad model that based on the words on the page, and not the community that created the words.

It’s not often as a Web performance consulatant and analyst that I find a book that is useful to so many clients. It’s much more rare to discover a book that can help most Web sites improve their response times and consistency in fewer than 140 pages.

Steve Souders’ High Performance Web Sites (O’Reilly, 2007 – Companion Site) captures the essence of one-side of the Web performance problem succinctly and efficiently, delivering a strong message to a group he classifies as front-end engineers. It is written in a way that can be understood by marketing, line-of-business, and technical teams. It is written in a manner designed to provoke discussions within an organization with the ultimate goal of improving Web performance

Once these discussion have started, there may some shock withing these very organizations. Not only with the ease with which these rules can be implemented, but by the realization that the fourteen rules in this book will only take you so far.

The 14 Rules

Web performance, in Souders’ world, can be greatly improved by applying his fourteen Web performance rules. For the record, the rules are:

Rule 1 – Make Fewer HTTP Requests
Rule 2 – Use a Content Delivery Network
Rule 3 – Add an Expires Header
Rule 4 – Gzip Components
Rule 5 – Put Stylesheets at the Top
Rule 6 – Put Scripts at the Bottom
Rule 7 – Avoid CSS Expressions
Rule 8 – Make JavaScript and CSS External
Rule 9 – Reduce DNS Lookups
Rule 10 – Minify JavaScript
Rule 11 – Avoid Redirects
Rule 12 – Remove Duplicate Scripts
Rule 13 – Configure ETags
Rule 14 – Make AJAX Cacheable

From the Companion Site [here]

These rules seem simple enough. And, in fact, most of them are easy to understand, and, in an increasingly complex technical world, easy to implement. In fact, the most fascinating thing about the lessons in this book, for the people who think about these things everyday, is that they are pieces of basic knowledge, tribal wisdom, that have been passed down for as long as the Web has existed.

Conceptually, the rules can be broken down to:

• Ask for fewer things
• Move stuff closer
• Make things smaller
• Make things less confusing

These four things are simple enough to understand, as they emphasize simplicity over complexity.

For Web site designers, these fourteen rules are critical to understanding how to drive better performance not only in existing Web sites, but in all of the sites developed in the future. They provide a vocabulary to those who are lost when discussions of Web performance occur. The fourteen rules show that Web performance can be improved, and that something can be done to make things better.

Beyond the 14 Steps

There is, however, a deeper, darker world beneath the fourteen rules. A world where complexity and interrelated components make change difficult to accomplish.

In a simple world, the fourteen rules will make a Web site faster. There is no doubt about that. They advocate for the reduction object size (for text objects), the location of content closer to the people requesting it (CDNs), and the optimization of code to accelerate the parsing and display of Web content in the browser.

Deep inside a Web site lives the presentation and application code, the guts that keep a site running. These layers, down below the waterline are responsible for the heavy lifting, the personalization of a bank account display, the retrieval of semantic search results, and the processing of complex, user-defined transactions. The data that is bounced inside a Web application flows through a myriad of network devices — firewalls, routers, switches, application proxies, etc — that can be as complex, if not more so, than the network complexity involved in delivering the content to the client.

It is fair to say that a modern Web site is the proverbial duck in a strong current.

The fourteen rules are lost down here beneath the Web layer. In these murky depths, far from the flash and glamor, parsing functions that are written poorly, database table without indices, internal networks that are poorly designed can all wreak havoc on a site that has taken all fourteen rules to heart.

When the content that is not directly controlled and managed by the Web site is added into this boiling stew, another layer of possible complexity and performance challenge appears. Third parties, CDNs, advertisers, helper applications all come from external sources that are relied on to have taken not only the fourteen rules to heart, but also to have considered how their data is created, presented, and delivered to the visitors to the Web site that appears to contain it.

Remember the Complexity

High Performance Web Sites is a volume (a pamphlet really) that delivers a simple message: there is something that can be done to improve the performance of a Web site. Souders’ fourteen rules capture the items that can be changed quickly, and at low-cost.

However, if you ask Steve Souders’ if this is all you need to do to have a fast, efficient, and reliable Web site, he should say no. The fourteen rules are an excellent start, as they handle a great deal of the visible disease that infects so many Web sites.

However, like the triathlete with an undiagnosed brain tumor, there is a lot more under the surface that needs to be addressed in order to deliver Web performance improvements that can be seen by all, and support rapid, scalable growth.

This is a book that must be read. Then deeper questions must be asked to ensure that the performance of the 90% of a Web site design not seen by visitors matches the 10% that is.