5 New Year’s Resolutions for Enterprise Architects

What are your architecture resolutions?

What are your architecture resolutions?

Happy New Year! Like just about everyone else, I make personal resolutions to lose weight, eat healthier, and get in more exercise so that I can be healthy enough to chase after my kid. I think there’s an architectural equivalent of these resolutions – things that we can do to prepare our companies for the changes coming in the enterprise. Here are five “foundational” components that I think are necessary to support the broader technology trends of cloud, mobile, social. I’ll probably elaborate more on these topics in the future. Maybe I’ll even take a resolution to do so. :)

1. Achieve World Class Integration Capabilities
You can’t predict how the market or even technology will change, but you can be assured that it will and the rate of change will increase exponentially. Companies that will get ahead are those that can adapt quickly to the change. Your business is going to want more capabilities, many of them delivered by small applications that need to be integrated with your processes, data, and applications. It’s imperative, therefore, to start thinking of applications and data as modular services that can flexibly be coupled together. With Gartner predicting that system integration will take up more than 50% of all IT spending by 2015, you can’t afford to tell your business that they can’t have the capabilities they want quickly because the system integration is going to take a year. If you haven’t started your transformation to a services-oriented approach to integration, it’s time to get going.

Some areas to start are: establishing an integration center of excellence, defining your integration technologies and SOA backplane, building your integration pattern reference architecture and governance. For the advanced shop, start looking at where cloud-based integration fits into your model.

2. Fix Your Identity and Access Management
Your business is working more closely with external partners, more applications are being delivered as SAAS, and your employees increasingly want to access their applications on their mobile devices without having to VPN into your network. To ensure security, you better have a good handle on identity and access processes, both internally and with external parties. This is an area where you need to have a good partnership with the business as they’ll own many of these processes and you will need to be able to offer a flexible model to support multiple use cases.

To start: get clarity on core IAM processes – how does the data flow to provision users and role changes? Have you identified and rationalized your authoritative identity sources and data model? Define your governance model with the business. Ensure that you have a standard solutions approach for on-premise and SAAS-based applications. For more mature shops, it’s time to start thinking about the roadmap to identity-as-a-service.

3. Putting the “I” Back in IT
In my view, IT’s central value proposition is providing relevant and credible data that enables insights into the business to drive action. A lot of the hype in information management in recent years has been “big data,” which while important for some businesses, has also been a distraction from basic information hygiene and other practices to make data useful. You don’t need big data to make a big difference in the enterprise – you do need clean, relevant, and timely data. This is another area where active participation with the business is a necessity, as they will need to steward and govern the data.

To get started: Identify areas where analytics can contribute to top line growth. Many organizations have also started with enterprise performance management, where multiple lines of business need to agree on definitions and relationships. Make sure you have good information architects to help mediate and articulate the data areas. Ensure that you have a well-tuned governance process. For more advanced organizations, big data is certainly an area to look at as well as data federation.

  4. Focus Your Security Strategy on What Matters
Security has been a hot topic for the past year, between the advanced persistent threats coming from governments foreign and domestic. Unsurprisingly, given the breadth of security gaps and risks, there are no lack of technologies to help:  intrusion detection, firewalls, anti-malware, SIEM, DLP, whitelisting, administrative passwords, digital rights management, encryption, unified threat management, and more. Given the myriad of technologies and vendors, this is an area that I anticipate will consolidate into a suite or managed service as there’s too much complexity and integration for enterprises to manage. This is an opportune time, therefore, to determine what aspects of the enterprise need to be secured and what is the right mix of technologies that deliver an acceptable risk profile.

To start: Identify what aspects of your data need to be secure and what is the acceptable level of risk. Don’t apply a one-size-fits-all approach otherwise you’ll be looking at a pretty expensive security bill. Try to minimize the number of technologies required to provide adequate security to lower your overall complexity.

5. Get Into the Cloud
I still run into a fair amount of resistance to the cloud – largely dealing with security and legal protections. These are fair concerns and there’s no question that the cloud model hasn’t worked through all of it… yet. That said, there’s no question that the cloud is inevitable, especially for commodity infrastructure or complex applications where the cost to manage is too high. I think vendors are racing to get the appropriate security and legal protections in place but there’s no reason not to start preparing your organization for it now.

To start: Start socializing the concept with your CISO – they’ll have a lot of concerns that need to get addressed and this is a good opportunity to partner and learn together to ensure that you have the right solution. A lot of early scenarios have focused on provisioning development and test environments in a cloud – I think this is compelling – especially from a cost perspective since your risk exposure is low. For more advanced shops, experiment with the public cloud and complex SAAS applications. For example, any third-party application that’s being developed using agile methods, I prefer to deploy as SAAS as there’s no way for an enterprise to reasonably keep up with that kind of upgrade schedule.

Two Weeks with Microsoft’s Surface RT

Surface

There's been a lot of buzz about Microsoft's entry into the computing hardware market with the Surface tablets. I've been using a Surface Tablet RT (the version using the ARM processor) for the past two weeks and have gotten a good impression of how effective this tablet is as an enterprise solution for knowledge workers.
 
Overall, I think Microsoft's first effort is quite good – especially on the hardware front. The tablet has a nice build quality, is sleek and lightweight, some very well-thought out design features such as the kickstand and keyboard covers. There is still room for improvement on the software front – the Surface RT is too heavily influenced by the PC-side of the tablet equation and not influenced enough by what Microsoft learned about mobility with their Windows Phone offering. For power enterprise users who want to use line-of-business applications or powerful productivity packages, I suspect you'll be better off with the more powerful Intel-based tablets. If your needs are simply focused on email, calendar, and the Office productivity suite (Word, PowerPoint, Excel), then you may find that this tablet is more than adequate for your needs.
 
Let me provide some context to my use requirements. I tend to spend a lot of time outside of my office, meeting with stakeholders in their locations – some of which are local and others involving travel across the US and in Asia. I've always liked the tablet form-factor and especially the characteristics that define tablet computing: lightweight, instant-on, and long battery life. My laptop, by contrast, is weighty and has short batterly life, necessitating carrying the power adapter (additional weight and bulk). This might suggest that I'm an ideal candidate for an ultra-book but I'd like to also maintain a touch-based interface to be able to do something quickly.
 
I've tried using an iPad as a laptop replacement but unfortunately, it's just not that easy. Apple scoffs at the notion of a tablet as a productivity device and it shows in their design. Without a fair amount of costly tweaks, it's hard to make your iPad into something you can do work on, not the least of which is the lack of native Office applications. 
 
HARDWARE
 
Microsoft has been raving about how much thought they placed into the hardware design and it shows. The tablet has a nice, sleek look – a magnesium alloy encasement with slightly beveled edges – only 0.37" thick. The dimensions will feel different for iPad users – the Surface RT is more of a rectangular shape than the iPad. Holding it from a portrait perspective will make it feel long and skinny compared to Apple's device. In landscape, it will look and feel quite natural. At 1.5 pounds, it's heavier than the iPad but I don't find it weighty. It has a good feel in the hand – it feels like a premium machine. It's certainly caught the attention of colleagues in meetings.
 
My favorite thing about the hardware is the inclusion of a USB port. It's only USB 2.0 but that alone makes this device far more useful, especially since I don't need to worry about carrying around some kind of dongle. I haven't used the port to attach to many devices but it worked just fine with my thumb drives and my other critical device – my mouse. I know what you're thinking – a mouse with a touch-enabled device? Yeah – absolutely. Touch is great for some things – scrolling, moving around large graphics, selecting large buttons but if you're trying to be productive, sometimes you need something with more precision. I've found this to be especially true on websites which are not mobile- or touch-optimized, for example, websites that have drop-down menus for their navigation. Zooming in- and out- to get that precision with my fat fingers is annoying.
 
The kickstand is as advertised. It's well-designed, innocuously fit into the back panel. A slight rift allows your fingers to ably pull it out to a comfortable and stable viewing angle. This is quite favorable to the iPad where I'm relying on a rubberized case which doesn't always conform to a stable shape and therefore is always at risk of toppling over with a bit of turbulence. 
 
TOUCH KEYBOARD VS. TYPE KEYBOARD VS ON-SCREEN
Since I'm evaluating the device, I also got the opportunity to work with both the Touch and the Type keyboard. Both work snap quite easily and flawlessly into the tablet body with the satisfying "click" that Microsoft has made a big deal about in the commercials. The Touch is thinner, with less weight than the Type (although the weights are quite negligible). They do feel quite different from one another.

As other reviews have noted, it takes a bit to get used to the Touch interface – it's completely flat and Microsoft wasn't kidding when it makes the claim that they innovated to figure out when you're actually trying to type something versus just resting your fingers. It's quite responsive and it doesn't take long to get the hang of it. I was typing with speed and accuracy within an hour of playing with it. That said, I did experience some anxiety every time I had to type in a password. Since I can't see the characters and there's no tactile response, I would get nervous when I needed to type in a special character. I found myself consciously holding down the shift key and pressing harder on the numbers to get the special character. Again, no problems with it – just takes getting used to.
 
By contrast, the on-screen keyboard lights up the key and gives a ticking sound to give more confidence. It would have been nice if Microsoft had at least put in an audio response to the keyboard.  The keys are a good size in both portrait and landscape although the landscape keyboard takes up quite a bit of real estate – nearly half the screen. In portrait mode, you can reach all of the keys pretty easily with your thumbs.
 
The Type keyboard works much as any other keyboard. It's has a good size and you have the immediate tactile response. I found that it was similar to just about any keyboard I've worked with. It does have a little more weight than the Touch and it's also quite a bit louder. Power users will probably find that they like this one better.
 
I didn't have a mini SD card handy to test and take advantage of the memory expansion slot. Much has been made of how Windows 8 RT takes up nearly half of the 32GB of memory in the model I'm using. Too early to see how fast I'd run through it but Microsoft is really pushing their cloud storage solution, SkyDrive, which gives you an additional 7GB for free.
 
I didn't run a battery test but I got through a full day of work without draining the whole battery. I'd guess that the iPad still probably lasts longer but the Surface does have a speedy charger. It's supposed to get a full charge in just two hours. 
 
SOFTWARE
 
The software is where I think Microsoft faces the biggest challenge on the device. In designing the software for Surface, Microsoft had a choice between letting it be influenced more by the PC side of Windows or more by the Windows Phone side of the equation. Unfortunately, it seems Microsoft chose the PC side.
 
Nowhere is this more obvious than in the Mail application. The initial mail application on Windows Phone 7 was criticized heavily because it didn't allow for consolidating multiple email accounts into a single inbox. The Windows Phone team corrected that in a subsequent update but clearly that message didn't get over to the Surface RT team whose Mail application still segregates all of your accounts. I also found the Mail application slightly buggy. Corporate Active Sync accounts had to be added in a certain order or else they wouldn't function correctly. Similarly, the Mail application wouldn't allow me to have both my corporate Active Sync account and my personal Outlook.com account on at the same time. Accounts that I deleted would sometimes be added back to the Mail application. Very frustrating.
 
It's unfortunate that Windows 8 RT didn't work more closely with the Windows Phone team – there might have been a possibility to enable the more than 70,000 applications already built for Windows Phone to get ported over – much like the iPhone apps can run on the iPad. Sadly, Microsoft has decided to start over which means they'll have at least three app stores: one for Windows 8 PC apps, one for Windows 8 RT, and one for Windows Phone 8. I've been told by some developers that it's not hard to port the code over for the three separate platforms but that remains to be seen. The app store so far is still pretty sparse.
 
Another disconnect potentially is that as you use the tablet, you learn that Windows 8 seems to have something of a split personality. The start screen and all the apps that you get from there are clearly optimized and designed for at touch experience, including the browser. Once you decide to use one of the Office applications, though, you'll discover there's a whole other side to your operating system that operates much more closely to the familiar Windows 7. For example, in this more Windows 7-like environment, there's another Internet Explorer browser which is more like the traditional IE and has a distinctly different interface than the mobile browser side. With some digging, you can also discover the more familiar Windows Control Panel and Windows Explorer dialogue windows. None of this is deal-killing but it does make the experience sometimes feel disjointed or awkward.
 
At the end of the day, you're running a full version of Microsoft Word, Excel, Powerpoint, and OneNote on a device that's a fraction of the size of a traditional laptop. I thought that perhaps Microsoft would narrow the functionality of these tools, but at least from what I can tell, these are complete versions of the software with no need to compromise functionality. The performance of the Office applications and the browser are quite speedy as is the OS in general. Where I saw the most performance lags were apps that needed to hook into the Internet – like Mail, People, Weather, etc. These apps generally had some clocking while they pulled down information from the Internet. Not sure why that's the case but it could get annoying waiting to see the most recent updates from your friends.
 
The other thing I like about the Windows device experience is the tie-in to your Microsoft Live account. Because Microsoft starts to sync nearly everything you do on your device with SkyDrive, it's really easy to get all your contacts, photos, music up and running on the tablet. Within minutes, I was able to get view of my pictures, see my friends' updates on Facebook, LinkedIn, and Twitter.
 
I did try out picture password and that seems to work out pretty well. It's a far more pleasant experience than the typed out passwords. 
 
BOTTOM LINE
 
I think this is a great device for individuals who don't have complex computing needs. The hardware is nice, delivers great performance and good battery life, and is a real winner with the addition of the USB port and full keyboard. I think the software is a bit awkward in a few places but it works – I'm sure it won't be long before Microsoft does the upgrade releases  to fix the bugs and bring functionality in line with the competition. If you're not using complex, enterprise software and don't have a complex security infrastructure to plug into, this is a great device. You get the basic productivity computing of a PC in the form factor of a tablet. What's not to love about that?

Voterball: The Data Disruption of Electoral Politics

 

If there's a great story coming out of the recent presidential election, it's how analytical, evidence-based methods are disrupting the conventional wisdom of political pundits and campaigns to deliver significantly more reliable forecasts and actionable insights.

 

The most visible story is around Nate Silver, the controversial former baseball statistician-turned political prognosticator whose FiveThirtyEight blog is published by the New York Times. Silver developed a statistical model that uses data to reject conventional political wisdom about what factors are important in an electoral race and accurately forecast election results. His predictions heading into the 2012 presidential race caused an uproar among pundits who denounced him as "a joke," a "one-term celebrity" "running a numbers racket," whose forecasts were "getting into silly land." For this year's presidential election, Silver's approach has been vindicated -accurately predicting the outcome in all fifty states and turning political journalism on its head.

 

Even more interesting than punditry, however, is how these data-based analytical methods are delivering impressive results for political campaigns. Time Magazine published a fascinating article this week about the Obama campaign's use of sophisticated data mining and analysis methods to not only drive an astonishing $1 billion in fundraising but also informed many other campaign activities such as election simulation, campaign ad-buying, and get-out-the-vote efforts.

 

One of the insights in the article was that President Obama's earlier2008 campaign had been unable to merge the data from their fundraising and voter databases, preventing them from having a complete picture of their support base.

 

“We analyzed very early that the problem in Democratic politics was you had databases all over the place,” said one of the officials. “None of them talked to each other.” So over the first 18 months, the campaign started over, creating a single massive system that could merge the information collected from pollsters, fundraisers, field workers and consumer databases as well as social-media and mobile contacts with the main Democratic voter files in the swing states."

 

They also brought in an analytics team five times the size of the previous campaign. As anyone familiar with information technology can attest, the above implies that the Obama campaign thought about the data that they wanted to measure, built a standard model by which the data could be integrated across a wide array of databases, applications, websites, mobile devices, stored in a data warehouse
where it could be mined for rich insights that served the campaign over and over again.

 

The notion that data has value is not a novel concept. That statement is obvious to most people in IT and business. I bring up this story of data-driven disruption to drive home the point that so many organizations pay no more than lip service to this concept. They make movements towards trying to clean up their data, trying to create governance, but at the end of the day, fail because they lack the resolve and commitment to make data a priority. I've seen this play out repeatedly in many organizations.

 

Too few global corporations insist upon and drive global consistency in their data models to break down the silo'd organizations, systems, and processes. Too few business organizations prioritize stewardship and the data cleanliness that would make that data useful for broader analytic purposes beyond basic operational reporting. Too many IT organizations focus on the technology component of their name, ignoring the information side. I think IT organizations and enterprise architects – especially those that want to be seen as a value-added partner instead of a cost center – need to transform themselves into advocates for stronger information management.

 

The story of Nate Silver, the Obama Campaign, and Billy Beane in baseball is about having an understanding and appreciation that it is the information that is an organization's most critical asset that needs investment and caring. Prioritizing that investment above all others helps to unlock the valuable insights about market levers, customer behaviors, actionable operational improvements, and accurate performance forecasting. And win elections.

Why the Yammer Acquisition Means Almost Nothing to Your Enterprise

 Yammer2
As soon as the rumors began that Microsoft was in talks to acquire enterprise social vendor Yammer, I started to get inquiries about it from stakeholders. When the acquisition became fact, some of my stakeholders expressed that Yammer had become a more viable option for how we could enable social collaboration within the company.

One of the main pitches that Yammer makes to enterprise customers is the ability to integrate with Sharepoint. They've spent a fair amount of energy building Sharepoint web parts that allow companies to expose Yammer feeds on internal Sharepoint sites and search results. Yammer's focus on Sharepoint was no doubt a major attraction for Micrsoft.

I would argue that while the acquisition is great for Microsoft, and absolutely fabulous for Yammer's investors, for most enterprises it's not really a net positive and potentially, could be quite negative depending on your company's disposition towards the cloud.

Yammer is a SAAS-product. It's highly unlikely that Microsoft will replatform it as an on-premise application. For companies that have enterprise agreements, this means that Yammer will likely be an additional license fee, just like Microsoft's other cloud offerings (O365, Sharepoint Online, InTune, etc.). At best, Microsoft might offer a discount for customers with enterprise licenses. If you're motivated to bring in Yammer to improve your Sharepoint environment, you don't need to wait for the acquisition to complete – you can purchase a license from Yammer today.

The net negative may depend on how comfortable your company is with having social collaboration tools in a SAAS environment on a public cloud and how much you trust Microsoft with this information. Without IT's permission, it's already possible for employees to establish Yammer accounts and network with others in the same company. Since the service is only very loosely tied to your directory service, there's risk that former employees and/or contractors could have access to the Yammer feeds since they need to be manually deactivated. That's a fair amount of overhead for an unsanctioned service.

A second negative – depending on how much you trust Microsoft and the SLAs that govern Yammer – is that those feeds are exposed to the company. The level of risk exposure could be minimal – an unsanctioned social network might not see too much traffic but it really depends on your comapny's risk threshold.

All in all, an interesting move by Microsoft to improve it's direction towards cloud-based offerings.

Five Implications of “All Architecture is Local”

Dominos
In my earlier post I argue that to provide value quickly, architecture needs to be thought of in the context of local needs more than enterprise needs. Here are five implications of architecting locally:

  1. Solve a local problem, not the enterprise problem. Making the resolution of a line-of-business problem instead of an enterprise solution will provide more defined scope, focus, and potentially better sponsorship to the project. This will also improve the credibility of the architecture practice in the long run – pointing to tangible success stories and building advocates among stakeholders.
  2. Pick the right local problem. Picking the right problem space to go after can be challenging – part of our criteria is to go after those areas where the business is already focused and where they are investing. We try to keep the business problem in focus rather than the solutions we're driving. (i.e. resolving account planning instead of Microsoft Dynamics CRM). In general, also tend to focus on areas where there is a heavy dependency on process or system integration. I think integration is an area where the value of architecture becomes more clear.
  3. Build a roadmap with shorter execution milestones. Not too many business stakeholders will have the patience to wait a few years for your grand enterprise solution. They need a roadmap that provides results much faster. 
  4. Forego elegant solutions for short-term results. Often the enterprise solution has a lot of components and can't operate effectively until they are all in place. Accept that some pieces are just not going to work well as you build out the capability. This may even mean having parts that are *shudder* manual. Be upfront about those limitations but be clear about the value that is being delivered in the short-term. 
  5. Design for modularity. While trying to deliver the right local outcome, bear in mind that the solution may need to scale for other businesses or the immediate business itself may change. Don't design too tightly to the local need – bring modularity and flexibility to the design so that downstream change will be easier.

 

This Week in Bad Architecture: What Floor is the AirTrain?

AirTrainFloor
Here's one of my favorite pictures of bad architecture that I use frequently in my presentations to non-architects. These pictures are from an elevator at Terminal 3 at JFK. Clearly there are at least three departments at JFK, each with their own naming standard for the floors. Rather than have someone settle the differences, the elevator (and unfortunately, the passengers) become the middleware for integration.

All Architecture is Local

500px-John_Donovaon_Tip_ONeill

My first career out of college was as a nonprofit lobbyist in Washington, DC. It was an education in many core principles of politics, including the famous saying from former House Speaker Tip O'Neill, "all politics is local." 

Speaker O'Neill coined this phrase to describe the lesson he learned from his first and only electoral defeat (running for Cambridge City Council as a senior at Boston College) "that a politician's success is directly tied to his ability to understand and influence the issues of his constituents. Politicians must appeal to the simple, mundane and everyday concerns of those who elect them into office. Those personal issues, rather than big and intangible ideas, are often what voters care most about."

In this respect, architecture is much the same. We have big and intangible ideas that are about driving value broadly yet many customers are focused primarily on their local concerns. For some, this means changing the language that we use to describe architecture – essentially taking lofty ideas and concepts and translating it into more tangible concepts that our stakeholders can digest.

For example, I don't use the word "artifact" to describe architecture work products – I use the term that's less arcane and more familiar – "deliverables." Tony Scott, MSIT's CIO, an former chief architect at General Motors says that he didn''t even use the word "architecture" with his clients because he was concerned that it would conjure up images of ivory towers instead of tangible business value.

I think Tip O'Neill's advice, however, goes beyond simply spinning a big or national idea into something more palatable at the local level. It's about making the interests of local constituents the focal point of the effort and using that lens to look at the big ideas. Most Members of Congress spend a considerable amount of their time in their home districts, listening to their constituents. They have district staff, whose job it is to help constituents with problems that they bring to their representative. The call or letter from a local constituent has far more weight in a Congressional office than non-constituents. Walk into the DC office of your Representative and see how the staff treats you when they find that you're a constituent. It matters a great deal to their success and so every good politician invests in their local constituency.

I've seen plenty of architecture practices lose sight of this as they pursue lofty, enterprise-scale solutions such as master data management, enterprise integration platforms, SOA, and data warehouses, to name a few. I'm probably not alone in having sat through too many Inmon versus Kimball debates.

I've worked on a number of platform projects where we looked at scenarios that ran across a gamut of organizations and tried to design a solution that would fit across a myriad of needs. While these projects succeeded, they were often challenged by lack of clear ownership/sponsorship, clients that were disengaged because we weren't focused on trying to solve just their problems, long delivery times,

At the end of the day, all architecture is local. Architects need credibility if they are to succeed.  If the architecture doesn't address the tangible need of a line of business with a problem, what use is it?