Blogs

RedMonk

Skip to content

What's in your stack?

As you’ve noticed, I started a new “series” of posts called “What’s in your stack?” I wanted to get a few out there as examples before I opened it up to submissions. I’ve always liked those “what’s in your bag,” “what’s on your desktop,” and similar peeks into people’s preferences like The Setup/usethis.com. So, I wanted to do the same thing, but for software development “stacks,” which I put in quotes because I mean not only the tools used to develop software, the frameworks in the software themselves, but also the process used to get the software out the door and then repeat the whole cycle over.

As you can see from the questionnaire (you should fill it out and tell me what’s in your stack!), there’s 4, core question: what’s your team do/who are you, what development process do you follow, what tools do you use to write the software, and an open ended question about any tools or techniques that recently worked really well, or didn’t. If you’re interested, check out the overview of how it works in the questionnaire.

The original idea was to ask 4 questions and then a follow-up question and just publish 1 or 2 of those Q&A’s. I’ve done that to a limited extent so far, but the answers all been so interesting that’s its unfair to keep them all to myself.

Thanks to Dan Kim for helping me flesh out the original idea.

Comments Off on What's in your stack?

Categories: What's in your stack?.

By cote
— May 26, 2011 at 2:18 pm

Cloud integration with JNBridge – Press Pass

I talk with the press frequently. They thankfully whack down my ramblings into concise quotes. For those who prefer to see more, I try to dump publish slightly polished up conversations I have with press into this category of posts: Press Pass. I spoke with John K. Waters and Paul Krill about JNBridge’s announcement today. Here is the press pass:

Do we need cloud integration?

[There are] now over 3,000 APIs on ProgrammableWeb…took 8 years to get to 1,000, 18 months to 2,000 and 9 months to 3,000
—John Musser at Gluecon, 25 May 2011

“As enterprise development teams start including cloud technologies in their applications, incompatible cloud platforms and APIs will be a huge road block,” said Michael Cote, analyst at RedMonk. “We’re already seeing a clamoring for tools and services that integrate this spaghetti bowl of end-points, and they’re only going to become more important to realizing the benefits of cloud development.”

There’s definitely a need to integrate existing Java and .Net based applications to cloud-deployed applications: be those SaaSes like Salesforce, or sales & marketing automation platforms, or custom built applications that companies are running on the cloud. While cloud is all the rage now, a huge amount of existing business IT – “enterprise applications,” if you will – are still running in the “traditional” way in on-premise datacenters. Many of those applications are, obviously, in ,Net and Java (both of which top the TIOBE index consistently, along with C). Rather than do some sort of whole-sale to-the-cloud refactoring, companies are going to need to find incremental ways to get the benefits of cloud – at least if they’re not lucky enough to make the dramatic changes needed to be pure cloud. And, there’s also just the simple need for companies to integrate their internal IT – applications, systems of record, accounting, etc. – with SaaSes.

Are these the same old issues just moved to the cloud? How big a problem is in now/in the future? How about the impact on developers?

Essentially, it’s the same issues, but with less time spent managing the end points, hopefully. Also, cloud-hosted end-points tend to be a lot more API friendly than traditional applications, if only by following simpler web services approaches. But, there is no universal fix for incompatible data standards in the cloud. True, there are de facto plumbing “standards” like JSON, REST, OAuth, and others that have emerged, and the favor that cloud applications tend to give dynamic, quick languages like JavaScript opens the possibility for faster and easier coding when it comes to doing integration work.

But that integration work – coding to get two end-points that previously didn’t talk to each other to, well, talk to each other – still needs to be done, and even more so when pre-cloud technologies like the Java and .Net worlds supported by JNBridge are taken into account. At the end of the day, there’s still no universal data formats that can be relied on to automate this passing data seamlessly between different systems: developers still need to put in work to make incompatible end-points talk to each other, even if cloud makes those end-points more “friendly.”

How well do you think JNBridge executes with this release on JNBridge CTO Wayne Citron’s vision of “cloud interoperability is any object or API, on any platform, in any language, anywhere”?

Well, there’s a whole world outside Java and .Net, the two technologies JNBridge is servicing here, but the Java and .Net world are a universe in themselves that needs plenty of help. There’s so much valuable data and process locked in Java and .Net applications that can’t just be left behind in whatever cloud-y future is out there – and refactoring all of that to be cloud friendly would be an onerous task. Instead, you need tools that help modernize those pools. Clearly there’s competition out there from

Clearly there’s competition out there from other outfits like MuleSoft who announced a cloud integration story yesterday, all the way up to big folks like Microsoft Cast Iron, Dell Boomi, and Microsoft. On the other hand, all of the interest in integrating between “the cloud” and “the ground” pretty well validates the need.

Disclosure: JNBridge is a client, as are Microsoft, MuleSoft, IBM, and Dell.

Comments Off on Cloud integration with JNBridge – Press Pass

Categories: Cloud, Press Pass, Programming.

Tags: , , , ,

By cote
— May 25, 2011 at 12:13 pm

Links for May 19th through May 25th

View from the office today

Disclosure: see the RedMonk client list for clients mentioned.

Comments Off on Links for May 19th through May 25th

Categories: Links.

By linkposter
— May 25, 2011 at 11:23 am

Alpha.gov.uk: What's in Your Stack?

For the third installment of What’s in Your Stack? I spoke with James Stewart, of Ket Lai, who’s been working on a team hoping to make the UK’s government’s public face on the web a better, more productive experience. This one is extensive, but it touches very nicely on both technologies used and the people management side of the project, including how it may fit into the governments “mainstream” process when the experimental phase is done:

First, describe company, the business that the stack is supporting

Last year the UK government commissioned Martha Lane Fox (founder of lastminute.com) to produce a report on government online. That report recommended a “digital by default” strategy where the first stop for citizens to access government services was through a single government website. Following on from that Tom Loosemoore (ex-BBC, Channel 4, etc) was asked to pull together a team and produce a public alpha of what that single government site could look like. Three months later, it’s live and we’re very actively soliciting feedback that will help shape the future of the government’s digital strategy.

How would you describe your development process?

Historically government has outsourced most of its development, but for this project it was seen as very important to have an in-house on-site team, following agile methodologies. Around 12 of us worked from an civil service office in Lambeth, completing weekly sprints. We worked from referrer and search logs to identify the 100 “needs” that most frequently brought people to government sites and we identified a few custom tools and some ways of packaging content to meet those needs. Getting the composition of the team right was obviously crucial for a project with a tight timeframe and high profile, but thankfully it really gelled. We were helped enormously in that by a fantastic project manager called Jamie who kept us all focussed and provided a huge amount of support.

In terms of the back-end development we tried to stay as close to the user needs as we could. This wasn’t the time to be building out platforms or making choices that we’d be bound to should the prototype be taken forward to be the government site. Our team was comfortable with Rails, Sinatra and Django, and so we used a combination of them along with some custom middleware that would support some of the geographic utilities we wanted, and that would ‘skin’ the content to provide a consistent look and feel for the site. Everything went through at least a couple of iterations, but there are a lot of rough edges.

Despite the whole team being committed to automated testing, the way we got started and a couple of architectural mis-steps early on meant we didn’t automate our testing as much as we should. That bit us a few times and it’s something I’ll be very keen to address in the future. Where we do have tests we are doing Continuous Integration, reporting into an IRC channel.

I was very pleased we were able to make use of ScraperWiki to gather a lot of the raw data we needed.

We’re deploying everything onto EC2 and amazon were very helpful with scaling advice and support. We’re managing configuration with Puppet, and it’s been great to have the freedom to experiment with server configurations as we’ve built the rest of our stack. And with the potential level of attention being a constant concern, Varnish has been a god-send, even if I am a bit disappointed that we couldn’t make version 2 work well with simultaneous mod_deflate and edge side includes.

What development (IDE, build tool, etc.) and project management tools did you guys use, if any? Given the dynamic languages, I’m curious what the team uses.

The front-end devs both worked in Textmate. We set up a VMWare VM for them with a full check out of all the code so they had a convenient environment to develop locally. We were really lucky that they’re both pretty comfortable with git so were able to fit right into the way the backend was developed.

The backend team was evenly split between Textmate and Vim. When working in dynamic languages we’ve not really found a need for much more.

All our code was hosted on github, using one of their “organisation” accounts. We started out using github issues for ticketing, but switched to lighthouse as we needed a bit more flexibility, and better ways to see issues across the project rather than just per-repository. Our project manager then used google docs to track progress, and we used a wiki to keep notes on work in progress, research done, and other documentation. He’s written a bit about the project management under the banner “agile does work in government.”

Over time, I’m curious how you guys will integrate all the cool, whiz-bang stuff into the staid, normal stack that I’d expect government to be more comfy with. To start, will the project transition to civil servants rather than the current mixed team?

Most of the team aren’t civil servants. A couple of people were seconded from other parts of government, but a deliberate effort was made to bring in “outsiders” to avoid being caught up in existing politics and long-running debates. Our Editorial Lead wrote a bit about that on the blog. That’s a pretty radical step for government which has long outsourced development work (though most departments have in-house teams for communications, editorial, etc). I can’t speak for any formal decisions that might have been made, but my understanding is that there’s a shift going on where at least some of the development work will be brought in-house and if this project is taken forward an internal team will eventually be picking it up. That’s partly about integrating it into the “normal way” but also about shifting that normal a bit.

In terms of the actual tech used, I get the impression there’s quite a bit of scope for continuing to use whichever tools are most practical for the job at hand. We’ve seen the US government adopting drupal quite widely over the past few months, WordPress has been used pretty extensively in UK government over the past few years, and there are some rails apps in production for government. So there’s definitely some momentum behind open source options and building on top of frameworks that can be built on and/or deployed rapidly and there are quite a few signals that we’re entering a period where there’ll be quite a bit of flexibility in technology choices and a lot more developer involvement in those decisions.

(If you’re willing to answer a few questions for a What’s in Your Stack? piece, feel free to simply fill out the new questionare, I’d really appreciate it!)

Disclosure: GitHub is a client.

rev="post-6742" 1 Comment

Categories: Programming, What's in your stack?.

By cote
— May 25, 2011 at 10:52 am

Cloud integration with Mule iON – Press Pass

I talk with the press frequently. They thankfully whack down my ramblings into concise quotes. For those who prefer to see more, I try to dump publish slightly polished up conversations I have with press into this category of posts: Press Pass. I emailed with Steve Wexler about today’s Mule iON beta announcement. Here’s the full press pass:

Q: What are the biggest strengths and weaknesses of this product?

What’s nice about MuleSoft is their experience and assets in this area. They started out life as a mature, cheaper alternative to other expensive, ponderous ESBs out there, for example going after Tibco. That kind of ESB and data integration work is dirty, tedious, and time consuming: successfully working with all the different end-points, applications, and data formats out there requires one thing, time, which MuleSoft has had a lot of. I expect to see that maturity brought to iOS, and hopefully the open source community that helps build out all that functionality as well.

Q: While MuleSoft is somewhat dismissive of Boomi and Cast Iron as outdated cloud integration solutions, they do have Dell and IBM attached to them, respectively. How big an issue is MuleSoft’s size for prospective customers?

It’s not easy competing against billion dollar companies. The other risky part, with respect to Boomi and Cast Iron, is that Dell and IBM really like those two offerings at the moment: the acquisitions haven’t gotten lost in those big companies. If you look through the cloud computing success stories of each of those companies, many of them rest on the cloud integration solutions Boomi and Cast Iron drive. As ever, MuleSoft’s advantage is being small, theoretically giving it the nimbleness to deliver broader functionality more frequently than the larger vendors without worrying about cannibalizing existing revenue. For example, with IBM you have all sorts of potential revenue conflict between Cast Iron’s possible ambitions with queues, integrations, and even ETL software over in their Info group. If you’re all into small vendor risk-management, as that type of MuleSoft customer you what you want to see is MuleSoft be so successful that a large company buys them and gives them the same attention given to Boomi and Cast Iron now.

Q: What are the biggest trends driving this market segment?

The category of cloud integration is seeing a lot of interest and, thus, use now. You can see MuleSoft moving to that demands here, while IBM Cast Iron, Dell Boomi, and dozens of other little startups and projects servicing that need as well. As with any type of software, whether it’s running on-premise or in the cloud, you usually need to use several applications in concert to get the most value. This can be simple things like integrating between CRM and marketing & automation software to start and keep up with what your customers are doing, to more complex things like syncing between your ERP system and cloud-based invoicing. While there’s definite savings and flexibility gains that using cloud-based platforms or, even just, SaaS applications, businesses will still need to look for ways to integrate data between all the end points on the cloud and on the ground.

Disclosure: MuleSoft, IBM, and Dell are clients.

rev="post-6731" 2 Comments

Categories: Cloud, Enterprise Software, Press Pass, Programming.

By cote
— May 24, 2011 at 2:30 pm

Service-now.com Becomes an IT Management Platform – Knowledge Trip Notes

The #know11 IT shaman

In the service desk field, Service-now.com quickly became an innovator and a leader in evolving the otherwise moribund category. First, Service-now.com deployed their service desk stack as a SaaS in the pre-cloud days, and now they’ve been grafting in learnings and innovations from the consumer space into the relatively staid area of managing IT processes, changes, tickets, and IT management in general. In addition to their ever impressive list of big name customer wins, their success is evidenced by how quickly incumbent vendors have SaaSed-up and started, frankly, taking Service-now.com a lot more seriously as a competitive threat. See IBM’s recent Tivoli Live service manager launch, for example. And with all the good stuff going at Service-now.com, that’s their biggest challenge: finally rating high enough on their competitors radar to defend the chasm.

Know11

I finally made it to their annual, US user conference, Knowledge11, this week. Aside from attending, I was there to give a presentation: Automating People Back Into IT, a round-up of the new technologies and practices I’ve been seeing, narrowed down to what IT can use to, well, do better. The attendees were nicely engaged and the announcements of upcoming features all looked good.

#know11

Here’s my trip notes:

  • Social as a feature – “Social” is a new, interesting area for Service-now.com. They’re deploying activity stream, Yammer-like micro-blogging functionality, and in-browser chatting (IM built and integrated into the tool). In general, they seemed focused on surveying the various consumer technologies in social and seeing how they can be applied to how the IT department interacts with the rest of the company (through the service desk, self-service portals, a service catalog, etc.), but also collaborates inside IT. Chris Dancy, of Pink Elephant, does an excellent job of painting this kind of vision, and to hear him tell it, Service-now.com has come a long way in a year from when he was poking and prodding on the topic.
  • PaaS – As with Salesforce.com, Service-now.com is in a good position to offer a PaaS. They actually do have a PaaS without making much fuss about it, or calling out their offering so directly to call it a PaaS. I’d expect them to explore amplifying this message over the next year. It’s a somewhat specialized PaaS: as Fred Luddy put it, if it involves forms and workflow, Service-now.com is a good fit. They’re not known for any particular expertise outside of the IT category, but a handful of customers have been building on-top of the platform for things like HR applications. The conference portal, which lists the schedule, allows users to gather around an activity stream and do micro-blogging, as another example, was built on-top of the Service-now.com platform. The venerable market-leader Remedy, of course, has a general purpose platform, ARS, at its heart – BMC has just chosen to limit its scope to IT.
  • Maturing the company – With their ongoing success, Service-now.com is obviously looking to “take the company to the next level” (IPO, buy-out, or other big exit, cash event). They’ve done the usual swapping in of a new CEO, replacing founder Luddy, and there were even a few financial analysts at the conference. There’s a general feel that the time is ripe for tech “startups” that’ve been forced to wait over the past years to finally exit…and with the likes of LinkedIn having a strong IPO, who can blame those hopes?
  • Automation – In a recent version, Service-now.com added on runbook and automation, the actual provisioning and configuration of IT devices. They’re just starting out here, developing it in-house instead of partnering or acquiring. Their ambitions here are growing as evidenced by the ongoing work there and the sessions on the topic. One session in particular was delightful: a service-now.com developer opened the floor for suggestions of things to automate and was genuinely surprised to hear three people ask for SAP. Just one person asked for mainframe, which brought a titter from the crowd. The thing to notice is the customer’s desire and trust to have Service-now.com configure such a core application. This is a crowded space, with some popular mavericks like Puppet Labs and Opscode, and it’s difficult to see it as an easy game for Service-now.com to win on their own.
  • In-house – a subtle, but important thing to notice is that Service-now.com very much prefers to do in-house development, making me wonder if they have NIH syndrome. For example, one of the announcements at Knowledge11 was that they’d replaced Lucene with their own search engine. I’m not a technical expert on the exact specifics, and no doubt a SaaS the scale of Service-now.com requires heavy lifting when it comes to search. Also, notice that their automation technology is in-housed versus partnering with someone else. On the partner angle, several would-be partners that I spoke with mentioned that Service-now.com was generally cool on deeper partnerships – again, my theory would be that the company prefers to develop in-house instead of graft on technologies. That’s a criticism Service-now.com rightly lodges at it’s incumbent competitors: a bunch of technology duct-taped together rather than organically developed as a whole.
  • Open with users – time and time again I noticed that Service-now.com employees were very open and engaging with their users, more so than I’m used to expecting from vendors. There were many sessions (such as the automation one above) where employees seemed genuinely interested in getting feedback rather than just broadcasting. The analyst relations they’ve done over the years has been similarly breezy, which I appreciate at.
  • Developer Relations – that porousness is a good asset going into something Service-now.com will quickly need: more developer outreach. It’s one thing to have service integrators (“SaaS integrators”?) helping customers get up and running, but Service-now.com is quickly moving into spaces where it would benefit from others building on-top of their platform: dealing with different devices and applications for automation; integrating with external social services; building out vertical applications on-top of it’s PaaS-we-don’t-call-a-PaaS. Here, Spiceworks seems to have proven out the value of mounting up developers (or, “people who write code” for those code writing admins out there) around a platform with their Extensions program which has 184 plugins at the moment.
  • JavaScript – a nice context for all of this is how much Service-now.com, apparently, relies on JavaScript, which is used both in the front-end and back-end. I’m starting to see that this mature JavaScript is a better way to provide extension points in your platform, which, coupled with the fact that developers as a whole seem better skilled and comfortible with JavaScript, lays the foundation for a pretty good developer-friendly platform.

What’s Next?

It’s really no big secret that Service-now.com is an important vendor in the IT Management space. Their ambitions to widen their scope beyond service desk (and “service management”) have been humming a bit, but are obviously in full-gear. They have a long way to go with automaton (it just takes time to get the code right), and their not-so-explicitly stated, but clearly existant, ambitions to expand into other enterprise application needs with a PaaS are young ‘n’ fresh as well.

The level of interest and engagement from their customer base, though, is a huge asset. Incumbent competitors (largely the Big 4 of BMC, CA, HP, and IBM) have quickly wised-up to the Service-now.com threat, however, and can be expected to burn through resources in a frenzied pace to catch-up and defend themselves…or buy the Service-now.com if feasible. If Service-now.com can master just one more “trick” – automation or LoB PaaS, more than likely – they’ll be an undeniably solid and broad part of the IT Management space.

Disclosure: Service-now.com is a client, as are HP, IBM, Spiceworks, Puppet Labs, and Opscode. See the RedMonk client list for others related to the above.

rev="post-6716" 3 Comments

Categories: Cloud, Conferences, Social Software, Systems Management, Trip Notes.

Tags: , , , , , , , , ,

By cote
— May 23, 2011 at 11:37 am

The Googlebot Wants Everything – #io2011 Trip Report

Day 1 Keynote

Last week at Google I/O 2011, the search giant unleashed itself into almost every nook and cranny of IT, consumer and “enterprise.” This positions them against pretty much every vendor out there, and with their billions in money from advertising, they’re like a wealthy but aloof giant that seems to have unlimited resources, just waiting to emerge from the haze of success.

Summary

Both Stephen and James have written up aspects of I/O – Android and Chromebooks. To pile on, here are my take-aways about Google from I/O this year.

The good:

  • App Engine matures, reminds world that it’s a 3 year old PaaS in this age of rising PaaS-mania.
  • Android@Home is an ambitious attempt to get Google into every part of your life, in your living room, as it were – and the laundry machine.
  • The Chromebook program is a bold attempt to provide desktops as a service and completely displace the classic Microsoft PC model, as well as Apple, of course.
  • Google is the epitome of “the consumerization of IT,” by which I don’t mean the easy idea of bringing an iPadAndroid tablet to work, but returning to a focus on end-users and delivering features at a cadence that matches consumer-think, enterprise desires be damned.

The bad:

  • Current bandwidth limits (and heinous over-charges) hobble most all these efforts.
  • Displacing Microsoft Office as the reason to buy the Microsoft platform seems an impossible task.
  • A two OS strategy with Android and Chrome creates two dev stacks, creating the classic problem of intra-company fragmentation that every platform company has.
  • Google’s notorious, release-early-kill-the-lame approach, along with their aloof nature at developer relations is odd, if just a “new” approach to platform “stability.”

App Engine

This three year old Platform-as-a-Service seems to have finally graduated, with a swath of new features and pricing updates. As they say, “App Engine now hostsmore than 200,000 active apps that serve over 1.5 billion site views daily.”

Awareness of App Engine has always been oddly low, almost providing a counter-example to the rising PaaS mania (which us RedMonk have as much as the vendors who weekly chest thump that part of the burger). Nonetheless, there’s a good stock of figures and customer cases for App Engine: the Royal Wedding (scale) and WebFilings (“enterprise – booga! booga!”), to name two. The platform may not do everything ever developer needs (that’s the well hidden Achilles heal of all PaaSes, by definition), but App Engine seems damn capability.

App Engine still suffers from lack of awareness and short-listing among developers I talk with. For example, when I sent the Royal Wedding case (built by Accenture, of all people!) over to one developer friend, more or less asking, “isn’t this enough?” they replies with “I guess so, but I think I would like to see a long term successful app that has grown strong. And I would like to hear from their developers about their satisfaction with using Google App Engine.”

Chromebooks

Google I/O

This is the most interesting of al I/O announcements: much like Ballmer decrying the iPhone as soon-to-be flop, the Cassandras are out in force. Their major beef is thick and juicy, though: you can’t beat Microsoft Office. Indeed, if Google things it can beat the Office lock, they’re setting themselves up for a nasty lyre contest: better to wait for the new Gods to come than to fight the old.

The Office hurdle aside, it’s a wonderful idea: client PCs and networking as a service, a subscription to a laptop. Cost it out though: $350-500 up-front for the hardware, than $20-28 a month with a three year commitment, plus a little more for 3G, plus $50/year for Google apps and whatever SaaS subscriptions you need. The range comes out to $1,119 to $1,737 see the correct pricing below in the comments…which ain’t too bad compared to theories on the rate for “traditional desktops.” (See pricing details over at Engadget.)

We all intuitively know that the issue is a cultural one: people wanting, mostly, Microsoft client software. The rest of the issue is building up the applications, both as SaaS delivered to the Chrome browser and Chrome plugins, if those are allowed. Developers are being given a massive incentive to be in the Chrome store: Google only does a 5% take on revenue. Compare that to Apple’s 30%. This app market place thing is always a chicken and egg problem: Google has to build a large enough mob of users waving dollars around to get developers to write to the Chrome platform. Google, however, has the cash to subsidize that: I’m pretty sure Rovio isn’t just making Angry Birds free in the Chrome store, as it is now, surely Google is paying them. Every other app store has this challenge – WP7, etc. -, going against Apple, and subsiding early developers seems the only route.

The demo-glow aside, I predict there’ll be a massive expectations mis-match when the Chromebooks come out in June (I’ll be getting one as an attendee, so we’ll see). Users are learning, thought, from the iPad and Facebook: it turns out you can get by with less from your computer, and you might even like it. Still, the Google Apps problem encountered in the City of LA case is illustrative. The thing there, thought, was just a need for training and realizing that Google Apps wasn’t a drop-in, 1:1 replacement for GroupWare. And nor should it be: the tools an information worker – any worker that uses IT, really – uses are changing, and they should.

The other issue is Google’s constant, self-inflicted bugbear: they’re aloof when it comes to customer service and long-term product management. As a corporation Google is still close to zero when it comes to building up that credibility: we’re a long way from “no one ever got fired boy buying Google.” (Though, maybe “using a free service from Google” is more appropriate). What people need to realize – and use for their risk/benefit procurement decisions – is how much that “genuflecting in front of your customers,” as IBM’s Steve Mills recently put it, costs in the end. A hell of lot more than $28/month.

If you think an iPad would work for you, the Chromebook looks viable. As ever, in these anti-cross platform times, the question is just getting the apps you want. And that’s all about developers, which Google is oddly good at, for now.

Aloof Developer Relations

In the grand scheme of Google, it’s good to explore the side about Google Aloofness mentioned above a little. Google doesn’t seem to have a unified developer relations playbook, strategy, or policy. Some teams are on it just fine, e.g., Android from what I can tell. Others are painful to see in action. Google is very friendly, across the board with developers and enjoys a good reputation (which should make you wonder, “why all the fuss here?”). There’s raw access to Google developers at I/O and elsewhere Googlers show up. But you don’t get the sense that there’s a policy for any of this: projects come and go (RIP Gears), external developers are left o sort out release trains, and what services and APIs are reliable enough to depend on at any given point in time.

One example from I/O typifies this. During one of the “fireside” sessions (come meet the developers and ask them questions), an attendee asked, “do you guys hang out in the IRC channel?” The Googlers at the front of the room had that painful silence only developers can muster up when they’re waiting for someone else to answer, they looked at each other murmured. Finally: “I think there’s one guy, who’s not here, who does… it’s probably a good idea.”

Most of these projects and services are all open source, or at least free, and you certainly get what you pay for. But upping the community management to at least hanging out – if not also being moderately responsive! – in the IRC channel seems like table stakes, esp. for “free as in the other part of the business makes billions each quarter.”

Frameworks like GWT have achieved phenomenal success despite this aloofness, but I don’t think other Google projects and services will be so successful until they become “products.” “Elite developers” are better positioned to put up with this kind of thing from the open source world, but the developer world as whole likes to know things like “Gears won’t suddenly be dropped with no clear plan for a replacement.” Google seems to be learning, though. I asked a question along these lines of some App Engine users: the Web Fillings folks said that it has been a challenge (I think he used a euphemism like “fun”), but that it works now. Indeed, as mentioned above, the App Engine team seems to be matured in these respects.

The Galaxy tab 1.0 – tablet vicissitude?

Finally, a short note on one of the items attendees received, a Samsung Galaxy tab 10.1. So far, I really like this device. I’ve been using it in place of my iPad (mostly for Kindle reading, calendar, etc.). The chronic, iPad competitor position exists, of course: it lacks some apps I’d like (OmniFocus, which I wager will never bust out of the Apple world), and can be a tad buggy at times (web pages have refused to scroll on two occasions and the clock is stuck in GMT, seriously). Overall, though, it seems better than the iPad. The hardware is excellent, it’s fast, and Android is only a little “weird lookin’.”

Trifling ticks aside – and ironed out – if Android continues on this path, iOS will finally have some serious competition. (We chat about it more on my non-RedMonk, highly unprofessional podcast DrunkAndRetired.com.)

Footnote

A Tweet from James this morning reminded me of something I left out: Google, like so many people, is using major parts of the Sun playbook. I have no idea if they know it or not, but so many of the strategies and “skate where the hockey puck will be” ideas are shockingly similar to the last days of Sun. In its last days (and before, actually), Sun was all over the idea of the Internet of Things and “ubiquitous computing,” cloud and advanced data-center design (using Siberian data-centers for better cooling was a frequent talking point), and they were even starting to get all consumer and end-user oriented with JavaFX and Java on TVs (!).

This pattern of using the Sun playbook comes up time and time again now-a-days, not just with Google: lots of IBM vision-ware is very Sun-like. After the first day of I/O I rhetorically asked Stephen O’Grady why it’d work this time, for Google. He took his time, but came up with a good answer the next day: Sun was always focused on the enterprise, Google is focused on the consumer. It’s a subtle shift, but it might just work this time.

Disclosure: Google is not a client, but see the RedMonk client list for related folks.

rev="post-6705" 4 Comments

Categories: Cloud, Conferences, Programming.

Tags: , , , , , , , ,

By cote
— May 20, 2011 at 4:42 pm

Growing Dell – #SSVE Trip Report

New acquisitions panel

Dell held an analyst event in Austin a couple weeks ago (titled “Services and Solutions for the Virtual Era”, or SSVE), coming out strong with its message of transformation and growth via an expansion in the general “enterprise IT” space. This was primarily based on two, traditionally non-Dell lines of business.

The Growth Plan

Our strategy around the efficient enterprise and flexible supply chain continues. We continue to develop and acquire key IP and enhance our sales capabilities. And we’re also narrowing our focus on three key solution domains, namely end user computing, data center and information management and services and all things cloud. Each of these solution domains represent key areas Dell has to win. If FY11 was largely about getting operationally fit, then FY12 is going to be about leveraging this position of health and strength to move more aggressively and accelerate our transformation as a services and solutions company. Customers are now seeing Dell in a fresh light and we’re heading into the new year with strength and optimism.
—Michael Dell, FY11 Q4 Earnings Call

The up-shot is that Dell is trying to grow beyond its “cheap boxes” niche, which of course it should be. Dell is trying to become like those other “elder companies” out there: IBM, HP, Oracle, and a handful of others that dominate in the space of hardware, software, and services. For Dell, this is the way out of all their problems. That problem, as my investing friends would tell me, is being perceived as a low to no growth company. Microsoft is in that spot as well – and the pitch-fork mob of growthies have recently been let loose on Redmonk more than usual.

After the two day event:

  • It’s still not entirely clear how Dell will first become a peer of and then win against the likes of IBM, HP, and Oracle (the tactics that is), but you almost don’t expect them to spill their playbook enough to make you feel like it’ll work.
  • What is clear is that, product-by-product, Dell has a lot to offer: they’ve got a software division on their hands if they can realize it and start running it as such.
  • Accelerating into this new role requires paying more attention to the practitioners involved: both the IT staff and the developers who’ll drive innovation and adoption. RedMonk, of course is always big on those “kingmakers.”
  • As a “one stop shop,” Dell is at the beginning. While Perot allows them to speak to health-care as a vertical (“applications” and “solutions,” if you will), they need to build up more of these “solutions” with actual software and skill. The needs are things like: building up the IT for new banks from scratch, saving telcos from being stupid networks, leaning up government IT, modernization efforts, and other massive IT changes.

The rest, from Dell, is actually pretty good. Dell can talk to being additive to acquisitions, esp. in storage, but also in servers where Dell has been good at innovating and delivering to new needs for boxes from traditional profiles to “custom” orders of 50,000+ units for high scale web shops.

Also, see the quick video interview Barton George did of my reaction on the first day.

Dell as ISV

Most of all, what I like is that Dell is building a software division. I don’t feel like they really think this exactly, but they should (and they’re certainly acting like it). Software is what will allow Dell to grow in the enterprise, giving their customers a reason to buy Dell systems (or “solutions,” if you like) rather than just boxes.

Dell needs to start thinking like a next generation ISV: something even beyond salesforce.com. What does that kind of software organization look like, how does it operate day-to-day, how does it innovate, how does it package technology, and then deliver it all to make the most profit? Rival HP needs the same thing, but Dell is in a great position: a clean-slate, more or less. Dell doesn’t have the shackles of success tying them to previous (software) revenue models and technology.

Luring Developers

Well, I think the largest cloud providers today, if you will, are those Web 2.0. Those are like new workloads. I think there’s very little of the traditional workloads that are being moved into that space today. And so, most of those — and so, that’s pretty much incremental opportunity. I know there’s a lot of concerns about at some point is there cannibalization and then what happens. We haven’t seen that. I think where companies are at on private cloud is very, very early. And the number of them moving a massive amount of their scale out into the public, very, very few. And so, most of the consumption there is social media and things that wasn’t even in the calculation. So, this is net new business.
—Brad Anderson, Wells Fargo Tech Transformation Summit

This is a hard slog, though, and just the beginning of the “fun” for Dell. The main thing they’re missing is a tighter focus and catering to practitioners, developers in particular (those “king makers” of tech companies). Developers love Dell monitors, but they need to love Dell software, hardware, and systems.

Mounting up a developers relation program here starts small, but requires patients and stamina. As with any marketing program, you have to offer something better, unique, if not cheap to draw a crowd. For example, Dell could create the ultimate build box: take one of the T7500 work-stations, load-up git, Jenkins/Hudson, Vagrant, etc. and do the work to make the development tool chain integrated and ready to use once the box was plugged in. And then there’s updates to the actual open source tools and configuration to make it take advantage of the 8+ cores on that box and the 12+ gigs of memory (that’s the base-profile I have, at least). Having a powerful build-box that development teams can plug in and have it just work would be attractive: just like the Google Search Appliance is (equipment OEM’ed from Dell, by the way). The build box could come with OpenStack, VMware, Eucalyptus (curated to simulate Amazon EC2, perhaps), or whatever cloud platform you wanted. Developers always want faster builds, and they’re not too shy about being impressed by over-powered hardware.

Here, the long-term plan for Dell is (a.) gaining relevancy with developers, and, (b.) as the developers need more hardware, they’re familiar with – even like! – Dell systems and hopefully buy it. On the extreme end, in what Dell would call the “Web 2.0” space (public web sites that use a lot of servers, cloud or not, to host their web apps), developers with datacenter needs might favor the systems platform they started with, Dell. Of course, there’s software, now, for Dell to sell as those teams accrue the needs of success: security, IT management, networking, storage, and whatever vertical offerings Dell can bring to the table. Throw in a few big-ass monitors as a bonus, and there’s even more incentive for developers to dive into a Dell build box.

IT Folks

There’s another group that Dell needs to cater to: sys admins, operators, IT staff, or whatever you like to call that lot. New technologies and practices – cloud, mobile as the new PC, SaaS (to call it out as its own force), and the vague but getting more fully baked “social” – are set to dramatically change the IT department, if not greatly reduce its importance. Who knows what the change will be, but it seems like it’s there.

Dell should be the hand-holder, thought-leader, or at least best-friend of admins: the company needs to help that community stay relevant and paid. Cost savings (or “efficiencies”) work well at first, but then IT needs to actually do something that contributes to their company’s revenue generation. Even if that contribution is just cost-cutting, there’s a question of how those savings are then strategically used. For example, as one Dell executive told me, Dell’s internal IT clean-up helped save $2B, which could be shifted, at the corporate level, to apply to acquisitions and the transformation Dell is going through.

A slightly less than impossible ambition

Dell Annual Analyst Conference

Dell has set itself up for the big, risky challenge of becoming a one-stop, enterprise IT shop. One of the “big boys,” as I put in my interview with Barton George. Dell needs to not only graft on and grow a software division, but build out their services arm with broad industry skills and programs. At the same time, Dell has to keep up its pose as the most affordable option that works. Normally, this challenge would be an operational quagmire (something Dell people have found themselves in too much in the past), but Dell is slightly better positioned then its peers: there’s little defendable reason to be loyal to the way things are/were, and almost no “legacy” models the company needs to cater to. In a very real sense, Dell has a clean-slate to start from, but enough existing assets to make their turn at the wheel of tech company transformation an interesting, if only slightly less than impossible ambition.

Disclosure: Dell is a client, as are IBM and HP. See the RedMonk client list for others mentioned or related above.

rev="post-6675" 7 Comments

Categories: Conferences, Enterprise Software, Ideas, Marketing, Quick Analysis, Systems Management.

Tags: ,

By cote
— May 19, 2011 at 11:39 am

Links for May 17th through May 18th

Disclosure: see the RedMonk client list for clients mentioned.

Comments Off on Links for May 17th through May 18th

Categories: Links.

By linkposter
— May 19, 2011 at 11:11 am

MindQuilt: What's in your stack?

In this new series “What’s in your stack?” I ask various development teams what technologies, processes, and practices they use to write their applications, get them out the door, and do the ongoing care-and-feeding required.

MindQuilt is often described as a “private Quora,” the Yammer of the all the Q&A sites out there. They’re a relatively new and small startup in Austin, but with a fair amount of use, deploying their software as a SaaS. As such, they’re a good example of he new type of application development and delivery that I see a lot of now-a-days.

I asked their CEO, Dan Kim a few questions about their stack and process:

What’s MindQuilt do?

I’m CEO of MindQuilt Inc, a knowledge management platform to track, organization, and capture new knowledge and documentation so your company can reduce project and operational risks.

What tools, languages, and platforms are you using for development and delivering your
software?

We use a python stack built on top of django. Mixpanel to track our funnel. We don’t watch people at all because we want to preserve people trust in putting their proprietary information on our platform. Instead, we use Olark so people can send in direct feedback or chat with us live if there is a problem.

Does python encourage spaghetti stacks? Is it more difficult to maintain long-term?

Yes it is a concern but that’s true of any language. Some languages (plus frameworks) combat spaghetti by forcing people into a rigid model but that tends to produce overly verbose code with lots of boiler plate. Additionally, it’s not spaghetti code that’s a problem but more spaghetti architecture. That’s outside of the scope of any language or framework. You need to do reviews often. No way around that.

What’s a tool that didn’t work out, and one that has worked surprisingly well?

We tried using Pivotal+Chat for a long long time. We found that it just didn’t work out so well for us. I think our team is too small. What has been working are daily calls between our US office and the technology team in Europe. If we need to share a screen we use GoToMeeting and screen-sharing.

What’s motivated you do off-shore development? What benefits are you getting?

We’re not really offshoring since:

  1. My CTO is Romanian and lives in Romania and I have worked with him in person for several years (in Germany and the UK).
  2. Our developers are full time employees with all benefits and stock options. They do nothing else but to work on MindQuilt.
  3. The primary reason for the dev staff to be in Romania is because that’s where my CTO is.

As for cost savings, yes we have a bit of cost savings over the market rate, but in a startup, you almost always have people working for under market rate because:

  • they are passionate about the product and have an equity stake
  • startups have no money

Disclosure: MindQuilt is a client.

Comments Off on MindQuilt: What's in your stack?

Categories: Programming, What's in your stack?.

Tags: , , ,

By cote
— May 19, 2011 at 10:57 am