Enterprise IT Watch Blog


March 18, 2010  1:07 PM

Performance Management is Critical for Agile IT

Guest Author Profile: Guest Author

This guest post is by Doug Willoughby, director of cloud computing for Compuware Corporation. Members interested in writing a guest post about an IT topic near and dear to their heart should e-mail community editor Michael Morisy at Michael@ITKnowledgeExchange.com.

Agile IT

Smart phones, social networks, Web 2.0, cloud computing, borderless applications: information technology is being reshaped by waves of disruptive innovations. Some enterprises will benefit from disruption, while others will be buried by it. Enterprises who position themselves to capitalize on innovation will benefit the most. To capitalize on innovation, successful enterprises are moving information technology to the forefront of product strategies, from a supporting role to the means of monetization.

In this environment, established IT organizations are likely to find their greatest challenge is their previous successes. To deliver on time and on budget, successful IT organizations have optimized their processes based on assumptions about the environment. Innovations and the changing role of IT throw many of these assumptions out the window. As a result, the primary challenge to established IT organizations is how to adapt their existing best-practices and tools to fit their new role.

One such best practice is the Agile development process. The Agile process has enabled IT organizations to be more responsive when faced with changing business requirements. In the past, the Agile process has been used exclusively by development teams. To address the operational requirements of ubiquitous access, social applications and cloud computing, successful IT organizations will become “Agile” across the application lifecycle-from requirements to operational deployment. These organizations will have all the advantages of being first to market, while enjoying lower operational costs.

To support an Agile application lifecycle, IT organizations need an integrated suite of tools that support each stage of the application lifecycle.

The Internet is the New Data Center

For the last fifty-years, IT has been optimized around the assumption that applications and data centers are fairly static. Lines of business owners generate requirements. Developers build and test applications in isolation before passing them over to the operations group for deployment. Once deployed, the operations group is responsible for monitoring the performance of the applications to detect and resolve problems.

Web applications are making the Internet the new enterprise data center. Research indicates that typical Web applications depend on six or more services located outside the direct control of the applications’ owners. These are not new-fangled cloud services, either. Rather, they are the bare minimum of ordinary Web services required to deliver consistent and compelling user experience, and include such Web staples as content distribution networks (CDN), ad distribution networks, content management servers, analytic services, streaming media services and other types of service delivery platforms.

The Web breaks many of the assumptions built into the traditional application lifecycle model. For example, thorough testing of many Web applications before their deployment may be a practical impossibility. There can be too many variables for proper coverage testing. Client browser compatibility (which release, what operating system, which plug-ins and what configuration options) is just the beginning. End user experiences will vary depending on where users are physically located–not just due to network latency and bandwidth, but also because the quality of service provided by Web services can vary by geography.

Performance Management is Critical for Agile IT

This is not saying that Web applications do not need to be tested before they are deployed. Instead, it suggests a new strategy that mirrors the Agile process that many development organizations have already adopted. This strategy is called the Agile application lifecycle.

The Agile application lifecycle is a cross-organizational approach that brings together line of business (LOB) owners, developers, and IT operations. Agile teams work closely with LOB owners to define requirement and IT operations to detect and resolve problems quickly. The approach is characterized by smaller, more frequent releases. New functionality is tested as extensively as practical, but greater reliance is placed on detecting and resolving problems once the Web application is deployed.

To be successful, Agile teams need integrated tools that provide seamless visibility from development to testing and on through to the management of deployments. LOB owners are concerned with end user experience, developers need end-to-end visibility into composite Web applications, and operation teams need to determine business impact and deep dive tools to resolve problems quickly.

Capitalizing on Innovation

Enterprises that see change as a normal business driver will benefit most from disruptive innovation. Adopting an Agile application lifecycle strategy enables these organizations to react quickly to change. Agile is a cross-organization approach, so teams need tools that encourage the integration of concerns from LOB to operations and enable them to move seamless from requirements to deployment. The end-to-end visibility and end user experience context provided by integrated application performance management tools, such as those offered by Compuware, are a critical component of any Agile application lifecycle strategy.

Doug Willoughby is currently the Director of Cloud Computing for Compuware, a leading provider of APM tools for Web applications. Prior to Compuware he was at Sun, which he joined in 1988 and where he participated in the development and marketing of some of Sun’s most pioneering and disruptive technologies, including Project Spring, Distribute Objects Everywhere (DOE), NextStep/OpenStep, and Java. Willoughby was also part of the team of 14 engineers and architects who developed “network.com,” Sun’s first utility computing offering

Compuware offers an integrated suite of application performance management tools. Gomez Actual User Experience XF and Vantage for End-User Experience provides LOB owners visibility into real user experience. Gomez Web Load and Performance Testing and Cross-Browser Testing tools, combined with Vantage for Java and .NET performance tools gives developers clear insight into how applications will perform when deployed. IT operations can leverage the full suite of Gomez and Vantage tools, including Vantage for Business Service Management, to understand the business impact of service problems.

March 17, 2010  8:30 AM

What secrets is your Intranet hiding?

Michael Morisy Michael Morisy Profile: Michael Morisy

That’s the question Intranet vendor ThoughtFarmer put to the world on Intranet Secrets, and the answers have been pouring in:

  • “Our intranet is optimized for Netscape Navigator 4.0.”
  • “I hate our intranet with a rage as white hot as the sun.”
  • “Just noticed that a hidden corner of our company intranet has a page with several lines marked “under construction” for, oh, 6 years or so.”
Ouch. And many of these are beautifully illustrated in the Post Secret mold:


March 16, 2010  8:51 AM

Windows Mobile 7: What is it good for?

Michael Morisy Michael Morisy Profile: Michael Morisy

More and more Windows Mobile 7 details are drip-dropping out daily, with the latest being that Microsoft’s taken a trick from Apple’s playbook and pushing application downloads through its own store, with a “forthcoming” enterprise option to centrally deploy corporate apps.

But will it even matter to the enterprise? Windows Mobile devices have long been a consumer also-ran, but have found die-hard users for specialized applications in areas from the warehouse to the road to extreme heat and cold situations. They often put the “computing” in mobile computing, but it doesn’t look like Microsoft is trying to keep that market with the latest update, which Ed Hardy dubbed the “Zune Phone.” It probably makes a lot of sense from Microsoft’s perspective, but what will power enterprise warehouse floors now? Will device manufacturers rejoice at getting to drop the “Microsoft tax” and go with custom Android builds? Will Microsoft have a legacy program to keep this market just sated enough not to leave? Will we see Windows Mobile 6, like it’s distant (distant!) cousin XP, be end-of-lifed for years and years to come as demand continues and companies resist upgrades?.

I’d love to hear your thoughts, either in the comments or at Michael@ITKnowledgeExchange.com: Does Windows 7 Mobile matter to your business, and if so, how — even if it just means you need to start planning a transition to something else?.


March 11, 2010  8:59 AM

The Wireless Office: Don’t Believe the Hype

Michael Morisy Michael Morisy Profile: Michael Morisy

While it might make me a public enemy in some circles, I have to stand by my assertion that we haven’t truly hit the age of the wireless office … yet. Josh Stephens, head geek blogger for SolarWinds, vehemently disagrees however, arguing it’s time to cut the cord.

He makes some decent points about wired vs. wireless security, although I’d argue that while .11n brings some marginal security improvements, it really ramps up the security complexity and opportunity for misconfiguration, which even he fingers as the culprit in Ethernet security lapses.

Where Josh really goes off the rail, however, is cost:

Next let’s debunk my buddy Michael’s point about cost. With 802.11n you can run 30-40 users per radio which means fewer expensive cable runs and fewer wireless switches. Combine that with mesh technology and you may not even need to run cable to all of the APs. This cost calculator from Aruba Wireless shows some great examples of how much you save. Mike, buddy, have you priced the costs of having an office wired with cat-6 lately? Even if you go all redneck, like I’m known to do, and run, terminate, and patch the cabling yourself it’s still incredibly expensive. In what universe is this not signicantly cheaper than a wired environment?

This is where he swallows the wireless vendors’ bait hook, line and sinker. A calculator, put out by a wireless vendor, shows that their products are cheaper than a competitive technology? Shocker! The problem is that, once you dive into the real numbers, you’re not going to stick 30 to 40 users on an AP, as Josh suggests. Sure, you could, and you could also use .11n to blanket a huge range, and provide throughput of 200 MBits/second, but it can’t do all those things as once, which a lot of calculators conveniently forget.

Another fact they conveniently forget: Even with .11n now an official standard, compatibility between sanctioned .11n devices is very imperfect when you’re talking about enterprise equipment. This means that no, you’re probably not going to save on all those pricey wiring costs because something, somewhere will need Ethernet.

But don’t take it from me: I tracked down Osaka Gas, hailed in 2007 by ComputerWorld’s Matt Hamblen as the largest “all-wireless” office, to get an update on what they’d learned and accomplished since becoming a poster child for cutting the cord.

There were some surprises:

1. This case study for the all-wireless office wasn’t then, and isn’t now, “all-wireless”!

Osaka Gas’ Toyoshi Matsumoto wrote to me:

We are using WAN for phones and PCs. Is that your definition of 100%? In fact we have NOT removed all LAN cables. Some fixed IP phones,  mainly used for receiving calls from customers or business partners are wired because the calls should be answered as the company not as an employee. Desktop PCs are also wired because they do not need mobility. Another use of wired LAN is the emergency use when WLAN gets unstable.

Three years as an all-wireless office, and a) It still makes more sense for some tethered IP phones b) Desktops are still wired and c) their WLAN still becomes unstable!

2. Toyoshi goes further, and says forget saving money simply by cutting the chord:

It’s not necessarily appropriate to suggest that wireless itself contributes to cost saving or improving efficiency. In many cases of  introducing wireless LAN as a replacement of wired LAN, you cannot expect cost savings.

Corner any wireless vendor for long enough on the cost savings issue, and they’ll invariably agree with this assessment. Sure, they have those nifty calculators Josh likes, but when you actually start computing the real totals, you’re maybe breaking even, but just as likely paying more for the Wi-Fi privilege.

3. Compatibility problems persist. Toyoshi puts a gentle spin on this one:

We have been using Meru since the project started. We have also other products from vendors such as Cisco, Aluba (sic), etc., to identify similarities and differences among them to determine their compatibility to our environment, because wireless technologies and standards continue to advance.

But the fact is, any major wireless deployment that’s dealing with more than students and their iPods has found the same compatibility issues. How often do you run into an Ethernet cord that isn’t compatibility with your laptop, projector, or other device (except those devices, naturally, that don’t have a port).

In the end, wireless is a great tool but it’s not the panacea that the industry makes it out to be. As Toyoshi put it, “‘Wireless’ is a means, but not a destination.” Don’t believe the hype otherwise.


March 11, 2010  1:42 AM

Checking back with 2006′s largest “all-wireless” office.

Michael Morisy Michael Morisy Profile: Michael Morisy

Editor’s Note: Below is the e-mail exchange between me and  Toyoshi Matsumoto of Osaka Gas Co., which was hailed almost four years ago as the largest all-wireless office. For background, see here. I reproduce the exchange below unedited. -Michael Morisy

1) How did the wireless experiment in 2006 go?

It went very well.  The project proceeded as originally scheduled and we achieved the 50% cost saving and improved efficiency as projected.

2) Did you continue on with Meru in this project?

We have been using Meru since the project started. We have also other products from vendors such as Cisco, Aluba (sic), etc., to identify similarities and differences among them to determine their compatibility to our environment, because wireless technologies and standards continue to advance.

3) Is this office now 100% wireless, or was it just deskphones that were wireless?

We are using WAN for phones and PCs. Is that your definition of 100%? In fact we have NOT removed all LAN cables. Some fixed IP phones,  mainly used for receiving calls from customers or business partners are wired because the calls should be answered as the company not as an employee. Desktop PCs are also wired because they do not need mobility. Another use of wired LAN is the emergency use when WLAN gets unstable.

4) Have you upgraded to .11n, and have you run into any challenges with that?

Not yet, but as stated above, we are trying various products and will try .11n in the near future because we are planning to replace the current wireless system in 2011 that is our Meru products’ economical end-of-life.

5) If you’ve moved to 100% wireless, why did you and what benefits did you see in terms of cost savings or efficiency?

6) If you’re not, what was behind the decision not to go 100% wireless?

It’s not necessarily appropriate to suggest that wireless itself contributes to cost saving or improving efficiency. In many cases of  introducing wireless LAN as a replacement of wired LAN, you cannot expect cost savings.

Our project started from the replacement of conventional PBXs. PBXs were so expensive that we could achieved 50% reduction of facility cost in terms of depreciation cost.

In terms of efficiency improvement, we not only introduced wireless LAN but also promoted to change our workstyle from paper-based one to full-digital one. Wireless environment and full-digital workstyle allow us to access, transmit and share real-time information anywhere, which leads to efficiency improvement.

7) Any other advice for companies considering going all wireless?

We believe that the use of wired and wireless systems and 3G and VoIP phones in optimal combinations is important for us to improve our work efficiency and thereby reducing operation costs.

“Wireless” is a means, but not a destination.


March 8, 2010  10:28 AM

Guide to Enterprise Mobile Communications and Productivity

Michael Morisy Michael Morisy Profile: Michael Morisy

This month, IT Knowledge Exchange is taking a special look at mobile communications and productivity in the business world. We’ll take a look at questions like whether it’s finally time we can cut the cord for office workers, who you should throw your lot in with during the mobility wars, and any other smart ideas for mobile discussions you e-mail in.

Frequently Asked Questions about Mobile Computing:

Still have unanswered questions? See what others are asking about cloud computing or ask your own IT question in our forums!

For a deeper dive, take a look at some of these excellent mobile computing book recommendations we’ve pulled together, or suggest your own:

Books on Mobile Computing:

Have another suggestion for this list? E-mail me at Michael@ITKnowledgeExchange.com or leave it in the comments.

Want to connect directly with experts? Read their blogs to hear straight from the horse’s mouth: The pioneers, cheerleaders and critics of the mobile computing landscape are often just a click away, and we’ve helped to organize the best of the best.

Top Mobility Expert Blogs:

What else would make this guide useful to you? Let me know in the comments or e-mail me directly at Michael@ITKnowledgeExchange.com with any additions, corrections or suggestions.


March 3, 2010  1:45 PM

Time to start picking sides in the mobility wars

Michael Morisy Michael Morisy Profile: Michael Morisy

While the tech world has been hyperventilating about the patent tiff  between HTC plus Google against Apple, a more serious war has been brewing in IT departments for years: How do you build a reasonable strategy based upon shifting standards and where future market share is nearly impossible to accurately predict? Just a few years ago, IT departments were scrambling to adopt policies for the iPhone. Now, it’s the Droid. And possibly the Pre. And the Nexus One.

Remember the good old days when everything was BlackBerry, except for some Windows Mobile thrown in for your industrial-strength computing needs? Now even WinMo is getting all spiffy, with Windows 7 making its mobile debut this Winter on a phone that doesn’t even include windowing.

So how are you picking sides in the mobile wars? Do you let your users pick their own phone, have a list of supported devices, or simply parcel out devices from on high? I’d love to hear your thoughts about what problems you’re struggling with, simply post any thoughts in the comments or e-mail me at Michael@ITKnowledgeExchange.com.-


March 3, 2010  11:11 AM

Don’t cut the cord yet: The truly wireless office is years away

Michael Morisy Michael Morisy Profile: Michael Morisy

I recently wrote about the paperless office and the post-e-mail era, two arguably worthwhile goals that share the fact that they’re nowhere close to reality except in a few small quirky progressive shops. The next myth I’d like to tackle is the wireless office: The dream of workers everywhere, but one we’ll have to wait on a bit before truly waking up to it.

The benefits of this supposed wireless wonderland, fueled by .11n and powered by pixie dust, is ease, accessibility and even cost savings. Imagine, users can just pony up to any office corral they want, pop onto the Wi-Fi, and even re-group with different colleagues with ease. No need to send an IT tech to reconfigure or troubleshoot their Ethernet jack, they just hop onto the same wireless connection no matter where they are. Those IT calls add up, and there’s a is a real cost. Several IT professionals estimated that each desk move runs a couple hundred dollars just in the time it takes to set the user up for access to the Internet and corporate network.

And the dream is nothing new: A few weeks after I started working at SearchNetworking.com in 2007, I was getting pitches about it. A quick search through my e-mail turned up tens of thousands of e-mails for “all wireless office,” from PR flacks, newsletters and mailing lists.

And flacks still tell me regularly: This is the time for the all-wireless office.

But the reality is, just like the paperless office and the e-mail-less office, we’ve got a long hard road ahead.

Take one of the top Google results for the topic, Matt Hamblen’s “All-wireless office launches for 6,000 users at Japanese company“.

Of course, once Matt dove into the details, he found that the office wasn’t really all wireless: A lot of workers still had traditional deskphones, and even the 6,000 wireless phone users had backups in case of a power outage.

But Michael, I hear you argue, that was three years ago. .11n has finally been ratified! Sure, and it’s brought a host of improvements, but in interview after interview with networking professionals, I’ve found that wireless companies routinely over-promise, under-deliver and leave customer after customer holding the bag.

Here’s three areas that give me pause:

Consistency: .11′s MIMO is a whole new ballgame for wireless, and the improvements are great. But they’re also overstated, because just as the greatest maximum reach has been extended, so has its unpredictability. That better-than-broadband throughput only comes through if you’re lucky, unless …

Cost: There’s a reason wireless vendors are so happy to pitch the all-wireless network: You’ll end up saturating your offices with access points. Access points that are “cheap” if they come in under $600. Access points that will have to be super-saturated once you discover that you have users clustering around in a meeting room. So much for saving on those desk moves.

Commodity: When you buy Ethernet cables, you pretty much buy Ethernet cables. And they’re a dirt chip commodity at this point. Forget that when it comes to wireless: You’ll have to pick a vendor and stick with them for a long, long time, bending over every upgrade cycle for whatever specific product they have to fit your need. And you’ll have to get the whole matching kit and caboodle, too: The controllers, the APs, the security consoles, the software that does your coverage mapping which, by the shows, shows you need even more of those proprietary controllers and APs.

I won’t argue that there’s not a lot of benefit that the ratified 802.11n standard brings, but be careful that your enterprise doesn’t get caught up in the spin.

If you’d like to read more, I highly recommend Lisa Phifer’s three-part series on .11n (registration maybe required), which advocates and explains a mixed network methodology, a compromise that, just like reduced paper offices, makes a lot of sense.


March 2, 2010  9:47 AM

There is no Knight in Shining Armor in Cloud Computing

Michael Morisy Guest Author Profile: Guest Author

Editor’s Note: This guest post is from Ed Laczynski, founder and CTO of LTech. -MM

I talk to a dozen or so IT managers, CIO’s, and decision makers every week.  My sales and business development staff speaks to another hundred or so.  Looking back at the past year of business in cloud computing, we’ve come up with an anecdotal trend.  IT management falls into two camps: pioneers and damsels in distress.

“Pioneers” are ready, willing, and able to take the plunge into cloud computing. They know that their job function might change, their users may go through some unrest, and their management will be tentative.  However, they also know that the world is changing.  The disruptive cost savings, productivity, and functionality that cloud computing provides is not worth fighting against; instead embracing it with the right mind towards managing, integrating, and operating the technology is the correct course of action.  The pioneers we worked with in the early parts of 2009 are already seeing the benefits.  Pride in ownership of successful deployments, savings of hundreds of thousands, if not millions of dollars, and control of their own destiny in this new era of technology give the pioneers the confidence to continue the adoption cycle of cloud.

However, “Damsels in distress” aren’t so happy about all this cloud hullaballoo.  They say cloud is really all about private datacenters.  Or, it’s all about some new product coming out soon from your typically enterprise software company, that promises to be a knight in shining armor – allowing them to keep the status quo and innovate at the same time.  These IT folks are content with managing hundreds of servers, in costly real-estate, with even costlier software licensing.   When opportunities to move to cloud computing come knocking, they yell “Get off my lawn!” and hug their servers and enterprise software licenses that surely keep them warm at night.

But business is changing.  It is about interconnectedness, collaboration, and scalability.  The very things the Internet was born from.   The companies that were built on the Internet – like Google, Amazon, and Salesforce, have built the framework for IT and it’s ready.

We are moving from an era when your typical desk worker received 30 emails per day, to one where hundreds would be considered a quiet day.   The tools of yesteryear – Notes, Exchange, Outlook, aren’t going to solve the problems of information overload that companies are being saddled with.  Massively multi-tenant systems like Amazon Web Services and Google Apps have stellar uptimes and reliability when compared with on-premise or pretend cloud solutions like Microsoft BPOS, which has had enough outages to scare away only the most dyed-in-the-wool Microsoft fan.   None of this is lost on sales and marketing departments, who want to innovate and keep a leg up and not be told they have to wait weeks for procurement of licenses and servers to stand up technology to win market share.

The knight in shining armor won’t show up because he doesn’t exist.  The last thing companies like Microsoft can do is cannibalize their on-premise server and productivity business with cheap, effective, cloud alternatives.  Even if they were willing to do it, can they really deploy the data center and scaling capabilities that their Internet-born competitors already have a decade jump start on?

The Pioneer is going to look at an IT problem, like standing up a development environment, or deploying world-class CRM, and laugh at its simplicity and cost.  The damsel in distress will wait for Topaz Cloud Connection and Scalability Server 2011 Plus, Enterprise Edition (for Servers) while the competition spends more money on sales and innovation than keeping those servers warm.  If you think the companies of yesterday are going to chew their own arms off to deliver you highly scalable, internet powered, cheap cloud computing, you’ll be waiting a very long time.  So think hard about what technologies and IT providers you want to cast your lot with in 2010.

The company (www.ltech.com) provides products and services that connect business to the cloud. LTech has successfully completed hundreds of cloud deployments and our cloud enablement products are being used by hundreds of thousands of users worldwide.

LTech is a Google Enterprise Partner and Amazon Web Services Solution Provider.


March 1, 2010  4:28 PM

A-Commerce: 10 Ways APIs will change IT Operations

Michael Morisy Guest Author Profile: Guest Author

Editor’s Note: Today’s guest post is by Sam Ramji, vice president of Sonoa Systems and former head of open source strategy for Microsoft. If you liked what you read, he has his own blog or you can follow him on Twitter. -MM

You’ve probably heard that Twitter’s API has been the primary driver for the fast growth and rapid morphing of Twitter’s service.  You may know that eBay and Salesforce.com get over 60% of their usage via APIs.  And in the last couple of months, you may have heard people at your company in marketing, business development, or software engineering talking about your own API.  If not, you will soon.

If you’re in the retail industry, this is going to make you very busy for the next few years.  APIs are a technology buzzword that basically equate to a new way to use the web.  In the 90s every retailer went “online” to take advantage of the cost of sales and margin improvements that came from having an e-commerce channel.  These sites enabled companies to “sell direct to millions of new customers”, and those who got online later had to race to catch up just to protect their businesses.

Now in the 2010′s there’s a new way to use the web – a-commerce, or commerce via APIs.  Mobile app and web app developers can use APIs to build very cool new applications that look and behave totally unlike your core website, but use your commerce engine just like a regular affiliate.  This lets them get to consumers who would never have come to your website, but love to use the app and therefore your company makes money.

While at first this may sound like nothing new, it turns out that there are a lot of new issues to manage.
The 10 New Factors of A-Commerce for IT Operations

1. Performance: API-driven demand patterns & load on infrastructure are really different from web-driven demand.  Developers will often wrap a database object directly in an API rather than shielding it with a web page that limits the number of rows that will be returned; programs will use that API in unpredictable ways that will load your system differently.  Added to that, many more new concurrent connections from thousands of new sources will be simultaneously hitting your backend servers.

2. Analytics: Channel sprawl is a good thing for margin, but tough on reporting.  There are multiple channels that affiliates are coming through – iPhone apps, tablets, web apps – and you’ll need to provide a combined view on their activity.  API traffic cannot be seen by Google Analytics or any existing web tool so you will need to figure this out.

3. Auditing: Recording the sources of the a-commerce transactions and integrating with affiliate management services to pay a-commerce partners is important.  Payment disputes will happen and you need to have a trail of data to show what happened in your systems.

4. Seasonality: Preparing for holiday rush is critical in order to run a trustworthy a-commerce service.  This requires not just performance forecasting and knowing what can be cached, but how to throttle low-value requests when high-value purchases are in the queue.

5. Security: The number of usernames and passwords are going to explode.  Don’t make users and developers build a new username and password to use your system.  By making OAuth the standard you can let users and developers log in using their Twitter or Facebook accounts.  This will save you a ton of hassle managing password resets and angry users.

6. Protection: Prioritizing traffic between web visitors and API users – who has priority when your infrastructure is under load?  Additionally, protecting against a-commerce threats requires filtering out XML header bombs, SQL injection attacks that come in via the API, and other new forms of attack.

7. Privacy: Ensuring that sensitive data isn’t exposed incorrectly requires knowing and controlling what customer and commerce data is leaving the firewall, staying in compliance, and ensuring PCI standards are met.  In an API world, this data is hidden in XML and JSON formats which you will need to scan and manage.

8. Evolution: Unlike a website which is under your control, or under the terms of “caveat emptor” when you are being webscraped, now there are affiliates who are depending on the API working a certain way.  When the development team changes their code and builds a new version of the API, you need to be prepared to manage apps that break.

9. Provability: SLAs multiply in this scenario.  Make sure that you can prove that your service was up and responding when upper management comes looking for who to blame when things go wrong for a high-priority a-commerce affiliate.

10. Debugging: this used to be something that just the internal development team handled by themselves; you may or may not have been involved.  Now there are a ton of new developers trying to figure out how to use your service, sending malformed requests, generating errors.

The specific combination of analytics, debugging, provability, and protection will come in extremely handy during the winter holiday season – being able to understand traffic spikes, identify misuse of your platform and removing that traffic while letting the good transactions continue to flow will be crucial in preventing downtime and maximizing revenue generating CPU cycles.

In the next articles in this series, we’ll dive deeper into each of the 10 issues listed above.  Let us know which ones you’re most interested in and we’ll cover those first!

Sam brings over 15 years of industry experience in enterprise software, product development, and open source strategy.  Prior to Sonoa, Ramji led open source strategy across Microsoft. He was a founding member of the AquaLogic product team and has built large-scale enterprise and Web-scale applications, leading the Ofoto engineering team through its acquisition by Kodak. Other experience includes hands-on development of client, client-server and distributed applications on Unix, Windows and Macintosh at companies ranging from Broderbund to Fair Isaac. Sam holds a Bachelor of Science degree in Cognitive Science from the University of California at San Diego, and is a member of the Institute for Generative Leadership.


Forgot Password

No problem! Submit your e-mail address below. We'll send you an e-mail containing your password.

Your password has been sent to: