The cloud applications tide has been rising for some time and it’s beginning to raise the boats in a spinoff IT sector – cloud application performance management.
A slew of vendors are popping up offering services to monitor EC2 spending, bandwidth, network performance, application performance and other advanced metrics. AppNeta, a Boston-based company located in the city’s growing “Innovation District,” is one of those vendors. Today, it released AppView Web, an addition to its PathView Cloud platform that will monitor app performance on the public cloud and offer insight into an apps performance over a network.
CEO Jim Melvin sees AppNeta as a way for users to increase the reliability of public cloud applications. He describes a common occurrence in IT departments – Salesforce.com is running slow in the office, IT calls Salesforce to ask why and they tell them that everything is running normally in the Salesforce data center. Melvin says this constitutes a typical call to AppNeta, which issues automated responses to IT through alerts, dashboards and reports that explain when the problem is on their side of the fence.
“If companies are looking to move towards public web applications, like Salesforce, like Google Docs, like a backup system like a mail system, they cannot take the plunge without some level of assurance, and that’s what we offer,” Melvin said.
Public clouds do have failures, as evidenced by recent outages in Amazon and Salesforce, but Melvin said that downtime in the office is often caused by network glitches or by inefficient use of resources. He describes two common examples – quality of service errors and BYOD.
Quality of service describes the way applications get prioritized on the network. Melvin says it is relatively common for a router to go down and be brought back up without anyone noticing, but that when new routers come online, they aren’t wired to respect quality of service, which creates a logjam.
The BYOD era adds an extra layer of confusion to the process. Anyone firing up their iPad in the office may inadvertently begin synching all their applications, from work-related apps to iTunes, soaking up a huge amount of bandwidth that takes away from apps in the cloud.
When asked what percentage of problems AppNeta respond to are caused by human error, Melvin estimated close to 90%.
“IT infrastructures are just incredibly resilient, they almost never completely fail because they are so dynamic,” Melvin said.
The market for performance management appears to be growing based on all the product news that has come out on an almost daily basis. Riverbed Technology announced a product upgrade that adds capabilities around quality of service today.
For more on application performance management (APM), check out contributor Crystal Bedell’s story on SearchSoftwareQuality.com.
Get ready, because it is coming. Developers are currently working through the kinks of its third beta, but iOS 6 will be here in the fall and enterprises are starting to take notice.
With roughly 200 new features in iOS6, including Apple’s highly publicized move away from Google Maps, is creating problems for developers who are frantically trying to figure the new world out before iOS6 goes live.
“So many of the developers use some form of GPS and mapping in their system, getting familiar with the Apple way of doing things is kind of a big deal,” Tom Lounibos said.
Lounibos is the CEO of SOASTA, a test automation company that is focused heavily in mobile app development. SOASTA’s new TouchTest platform, which allows for test automation on a large number of devices and accounts for new user interfaces like Jester, is designed to increase the availability of mobile test. He cited an IDC report that stated 80% of testing is done manually in mobile, mainly because there are so many different devices.
“Most of our customers have had to do manual testing, and they’re getting to the point where there are too many points of failure to keep up with,” he said.
From his perspective, the 6.3 beta of iOS is looking to be more stable than previous releases and the new features, especially the visualization and mapping ones, have developers in his circle excited.
“The user experience is changing. You used to have one-dimensional cool stuff like maps. Now instead of looking at a map while you’re driving, you can talk to Siri about that map,” he said, adding that he thinks this is just the beginning. He cited machine-to-machine interfaces as a rapidly changing segment that will forever alter the application world.
He also speculated that a world where applications are no longer tied to devices is soon at hand. He envisions things as futuristic as map applications popping up on car windows as being within reach.
“We’re just at the beginning of some enormous changes in the way people will interact with applications,” he said.
Lounibos had some other thoughts on the mobile app world and the issues facing it.
On Windows’ entry into mobile
“The question mark has been out there and with RIM being in whatever limbo they have been in, it’s caused a big opening for someone else to come in and be number three.”
“I think Microsoft is making some good news. I think Windows has a big following and a big developer community already in place; I think the tablet announcement has been big.”
APIs and the companies putting them out
“A lot of brand names where you would say ‘Are they still in business?’–all of sudden you see their API strategy. You’re going to see a lot of old-school brands jumping the shark.”
On cloud computing
“I really believe cloud computing is a byproduct of mobile, and not the other way around.”
- The mobile development market, meaning build/test/deploy/manage or mobile PaaS, is looking to be a $100 billion market.
- Believes that mobile apps will hit 2 million by the end of the year, up from an estimated 800,000 apps currently.
- Predicts that the most growth in mobile apps will be in enterprise mobile because “they jumped in late.
– Adam Riglian, @AdamRiglian
What do Walgreens and APIs have in common? Ubiquity.
The pharmacy giant partnered with Palo Alto, Calif.-based API management company Apigee to launch an open API around its QuickPrints service, allowing third-party developers to incorporate Walgreens photo printing service into their own photo-sharing applications. The new program launched with five partners – Pinwheel, GroupShot, Kicksend, Pic Stitch and StillShot.
“Really what it’s showing is that businesses of all types need new ways to reach consumers and new ways to scale,” said Marcia Kaufman, principal analyst with Needham, Mass.-based IT analyst group Hurwitz & Associates. “The API is really no longer just for an [online] company, it’s not just for Google.”
Walgreens initially launched QuickPrints as an app, but decided to also offer the QuickPrints SDK (software development kit) so any photo sharing service could include it. Third-party apps that use QuickPrints get a share of revenues after photos are purchased. For users of third-party apps supporting QuickPrints, they will now be able to select a photo for printing on their phone and pick it up an hour later at a Walgreens store, with no need to create a separate Walgreens account or login.
“We always had the vision of offering that functionality to third-party application developers,” said Joe Rago, senior product manager at Walgreens.
Rago said that this is the first step into the API world for Walgreens, but that it wouldn’t be the last. Taking advantage of the enormous digital photo market was his team’s first goal, but that eventually they plan on finding other ways to use APIs to expand Walgreens business.
“A good way of putting it is putting an API around our stores, bringing our stores to the internet,” he said. “Photo was a logical first entry point, just with the sheer popularity [of it].”
Kaufman compared Walgreens’ open API to the pharmacy chains inclusion of on-premise photo processing. She said that it was innovative at the time for a pharmacy to become a one-stop shop for birthday cards, groceries and photos, and added that this is no different.
“It puts Walgreens in a very innovative position, as opposed to closing up shop and saying this business of producing photographs isn’t so lucrative anymore,” she said.
In “Vendors aim to get cloud app development off the ground,” Redmonk analyst Stephen O’Grady said that cloud IDEs share advantages over desktop models, but that “the reverse is equally true.”
Cloud9 took another step towards closing the gap between what’s offered offline and online with a new release announced today. The latest upgrades focus on beefing up collaboration capabilities, adding autocomplete, additional language support (Ruby, PHP, Python) and the ability to work offline.
Autocompletion was the most requested addition from the user community according to Cloud9 CEO Reuben Daniels. The ability to autocomplete was added for Node.js, the language most used among Cloud9 developers.
“Developers don’t have to remember the API calls anymore,” Daniels said. “They can just type dot and it will appear.”
If autocomplete was the biggest perk for the current crop of developers, the ability to work offline is the upgrade most likely to draw in desktop users. Daniels said the addition was aimed more at startup users than customers in traditional enterprise environments.
“The enterprises that use us don’t have any problem with that, because the developers there just run inside their network and that’s fine with them,” he said. “We’re catering to a lot of startups, people that are younger and still have the ability to work wherever and whenever they want to. For these people, that type of flexibility is very important.”
Cloud9 is also attempting to recreate the desktop IDE experience by adding workspaces, secure environments separate from other development environments. In each environment, developers have their own container for files, the ability to run apps and process isolation. It’s something that would be taken for granted on a desktop, but is an important step for the cloud, according to Daniels.
Offline support, collaboration and autocompletion are available to all users, while the workspace in the cloud additions are premium features.
There is no predicted time table on how cloud and desktop IDEs will compete with each other in the future, but with the latest Cloud9 release it appears cloud IDEs are trying to take away reasons why developers would choose desktop options.
– Adam Riglian
Mirantis co-founder Boris Renski endured my clumsiness with the camera tripod and microphone setup to deliver a compelling conversation on his company’s OpenStack evangelizing and partnership with Dell.
Mirantis is an OpenStack engineering service company, but what sets it apart from some other professional services organizations is that they don’t believe in value added offerings.
Renski made it a point to note that Mirantis isn’t interested in making a product out of its OpenStack expertise. He plans on sticking to the “vanilla” distribution of OpenStack and focusing instead on conquering the biggest barrier he sees to OpenStack adoption – it’s moving too fast.
He says that some enterprises he works with are running in production on OpenStack Cactus, while OpenStack is now several releases ahead with Essex and another new release, Folsom, is not too far down the road.
To deal with the speed of this, Renski believes that Mirantis is best suited in guiding enterprises through the updating and upgrading process in a non-disruptive way by doing things incrementally. The approach is called infrastructure as code, which he likened to devops for infrastructure.
Mirantis partnered with Dell on developing Crowbar, a management tool that helps users deploy OpenStack.
“This is a vision we very much share with Dell,” Renski said.
It would be remiss not to mention Renski’s sense of humor. Moscow-born, Renski has been wearing a t-shirt that says “In Soviet Russia, cloud deploy you!” at the conference.
Stay tuned and follow me on Twitter @AdamRiglian.
Day one of the OpenStack Conference got off to a flyer of a start with a pair of keynotes both provocative and enlightening.
First to the mic was Chris Kemp, OpenStack co-founder and co-founder and CEO of Nebula. His points were simple – OpenStack is a stack, not a product. He pounded on that theme throughout his keynote.
To that end, he explained that the idea of OpenStack is that of a cloud ecosystem designed to be an open space for partners to collaborate.
Kemp argued against a recent report from Gartner analyst Lydia Leong, stating that he did not think that OpenStack competed with VMware or Amazon Web Services. After his speech, he repeated the point in a tweet, saying “Today’s #OpenStack, VMWare, & AWS ecosystems are (and should be) complimentary – each are optimized for different applications”
In his speech, he said that VMware was solving a very different problem than OpenStack. Kemp said that VMware has done a good job taking 25 years of legacy software and making it run in a static environment.
“Enterprise applications don’t like agile and dynamic. They are designed to be static,” Kemp said.
He added that while Amazon’s implementation was innovative, that the steps it took into cloud computing were “evolutionary, not revolutionary.”
Following Kemp was Zorawar “Biri” Singh, senior vice president and general manager of HP Cloud Services. Singh gave an overview of HP’s cloud strategy, discussed its partners and announced that public beta would begin on May 10.
Singh said that HP was active in OpenStack’s project policy board and that the company was open and ready to learn from its consumers.
He spoke on the “paramount” importance of APIs in a world where communicating between clouds will become critical and said that HP was coming from a point of view that, “standing up VMs” is a “2009 phenomenon.”
I was able to sit down with Singh for a one-on-one after his keynote and will have more on HP later.
Stay tuned and be sure to follow me on Twitter @AdamRiglian.
Private beta users gave cloud infrastructure management tool OpDemand plenty to do before its general availability launch on Monday. Integrate with social coding repository GitHub and give us more to look at on the user interface dashboard, they demanded.
OpDemand complied. Chief Technology officer Gabriel Monroy says that OpDemand’s platform is now “tightly integrated” with GitHub, offering single-click signup for GitHub users and other improvements that make it “as easy as possible to go from GitHub into the cloud.” He adds that the UI has been upgraded to a real-time interface that gives users a detailed look at what is going in in the guts of the system with little latency.
“We call OpDemand platforms platforms for a reason,” Monroy said. “They are application-centric and offer users the benefits of a PaaS but with control of managing your own infrastructure. Really it’s about masking complexity and making things simple for users without taking away control.”
Last year, SearchSOA.com profiled an online art gallery that used OpDemand to reduce IT costs by 40%. The galleries vice president of engineering cited OpDemand’s simple interface, easy configuration and infrastructure templates as the reason he picked the startup over more established competitors.
Monroy said that OpDemand, which currently supports Amazon but is planning on expanding to other clouds, has grown its template library to the point where it now incorporates things beyond just managing servers.
He cited elastic load balancing, virtual networking infrastructure, relational database as a service and autoscaling as Amazon offerings that don’t fit into an often server-centric view of cloud computing.
“None of those things are servers but they are essential to working in the cloud,” Monroy said.
He added that they have also beefed up configuration management, what Monroy calls the “future of system administration.” Integrating with Puppet Labs, an IT service automation software designed for system administrators, users are now able to apply policies across all systems rather than logging into them individually to make changes.
“It’s really a necessity when you’re managing complex environments in the cloud,” he said.
As far as the future, Monroy expects the next step for OpDemand will be expanding to other public clouds, citing OpenStack as a likely next step, as well as mentioning the emergence of HP into the market.
“We chose Amazon at this point because they offer the best technology, the most stable and mature IaaS offering, but that’s beginning to change,” he said.
Hewlett-Packard is repositioning itself in the cloud market today with the announcement of a series of improvements, upgrades and new products under its Converged Cloud banner.
If the move is anything, it’s comprehensive. HP makes it clear that it’s trying to accommodate all comers into the cloud market and is willing to guide them through any cloud implementation — public, private or hybrid. Initial reactions from some analysts suggest HP is late to the party when it comes to cloud and that this move represents another attempt to close the gap with competitors.
“HP is playing catch-up regarding its overall cloud strategies and solutions, including its new Service Virtualization 2.0 and private cloud management capabilities, and how it is working with partners to build private clouds for its customers,” said Jeff Kaplan, managing director of Wellesley, Mass., consultancy THINKstrategies.
Service Virtualization 2.0 is the most crucial component of the new package for developers. With it, HP has created a testing environment for developers that allows them to test from the get-go, a system upgrade that accounts for the real need for agility in most enterprises.
“It fundamentally provides the glue necessary for development teams to create their own clouds,” said Matt Morgan, senior director of product marketing for applications at HP. “You can re-establish business as usual for a software development shop. They can build a plan that lets them test on day one.”
Morgan touted 25 new features in Service Virtualization 2.0, with some of the highlights being a full RST stack, XML protocol capability, protocol stacks with multiple end points, Ajax, the ability to classify response times to change response rates, server management capability and a new user interface.
He also cited new application lifecycle intelligence technology as a key selling point for developers. Through one platform, developers can perform ALM tasks while engaging others using new social collaboration tools, which include new dashboards and mobile and tablet apps.
“At the end of the day, everything is about agility, and the social collaboration is just a facilitated agility,” Morgan said.
– Adam Riglian