Category: software

Amazon’s “state of the cloud” — it isn’t stacking up

by Cortney Fielding

Werner Vogels - CTO, Amazon.com - Structure 2011

When it comes to the cloud, just forget about the stack. Cloud computing is evolving beyond the data structure of interchangeable layers, said Werner Vogels, Chief Technology Officer and Vice President at Amazon, during his “state of the cloud” address at GigaOm’s Structure conference.

“I’m moving away from this picture of a stack with infrastructure on the bottom and platform on top. It’s outdated,” he said. “Most of the Amazon cloud is much more than services.”

Instead, the cloud — and Amazon Web Services — is about pulling together services from a million different ecosystems that allow enterprise customers to build highly sophisticated applications. And while much of the enterprise world says they are buying instead of building, they are actually building too, he says. They just aren’t starting from scratch.

Both Oracle and SAP are now certified for software production on the AWS cloud, allowing businesses to build their own very sophisticated solutions using “standard, off the shelf” pieces from across different providers, he said.

Vogels was on hand to discuss where Amazon sees cloud computing and where he plans to take the industry next. Vogels helped elevate the cloud computing industry into the mainstream when his team at Amazon launched EC2 and S3.

The Amazon public cloud accounts for an estimated 40 percent of the total public cloud market. And Vogels said the number of objects on AWS has more than doubled in the last year, from 150 billion to over 339 billion objects. And that’s just the beginning. “On a daily basis we are adding as much capacity as Amazon retailer used in year 2000,” he said.

Looking back to the early days of AWS, Vogels said the original mission was to enable businesses and developers to build scalable, sophisticated applications. So how did that pan out?

“Predicting the world of now has been really hard. But if I look back some has come through,” he said. “Both developers and businesses are building increasingly sophisticated applications in the cloud.” He said the cloud has allowed young businesses to become extremely fault tolerant.

Vogel also pointed to expanding geographies in Amazon’s cloud network. AWS now operates in five regions and recently launched in the Singapore and Tokyo regions.

He said within the next year, AWS will focus on making governance, billing and user-management easier, as well as offering customers “more detail and more choice.”

Asked to make a ten-year prediction for the cloud, Vogels said it just wasn’t possible.

“Almost on a yearly basis we see ‘cloud’ evolving more. We definitely see customers building more and more sophisticated applications,” he said. “As such it is still day one. There is a lot of work to be done.”

 

The future of the cloud, as seen by Cisco

by Katie Fehrenbacher

Lew Tucker - VP and CTO, Cloud Computing, Cisco - Structure 2011

 

 

 

 

 

 

 

 

The future of the cloud is a programmable network with applications that are smarter about using resources, said Lew Tucker, VP and CTO of Cloud Computing for Cisco at Structure 2011 in San Francisco on Wednesday. The industry needs to be able to provide more access to applications and be able to “program the network how we’d like to,” said Tucker, noting that this was one of the driving factors for why Cisco has embraced OpenStack, an open source cloud project pushing an open cloud computing platform.

This isn’t the first time that Tucker has rallied for the importance of developers, even when it comes to cloud network architecture (in contrast to the more obvious developer fodder like mobile apps). Cisco’s latest move to work with developers and the open source cloud is its Network as a Service (NaaS) project, which would give OpenStack developers the ability to easily create and manage virtual networks using simple APIs.

Take that a step further and Tucker envisions a world in the future (what interviewer Joe Weinman called his “crazy town idea”) where cloud applications are “self aware,” when it comes to the resources they are using, like enterprise resources and power utilization. Elasticity — being able to scale up and down resource use — will be an important function to moving the network to the cloud (as well as a way to make the cloud more energy efficient). Eventually “applications will become more intelligent about using resources.”

Tucker also noted that the future of the cloud will be “dominated by data.” It’s going to be easier to move applications around than large data sets, like in industries like genomics. Finally, Tucker thinks that the whole debate about public clouds versus private clouds will eventually go away.

 

Why iCloud Will be as Important as the iPod

By Adam Richardson

Apple’s World Wide Developers Conference keynote last week will be
remembered for two things: the bloodbath of disrupted developers and
apps it left in its wake, and that it was as important for cloud
services as the iPod was for digital music, and that the iPhone was for
smartphones.

The Developer Bloodbath

Despite the many cheers from the crowd of developers at the keynote, I
reckon there were several hundred third party developers and apps
collectively put on notice (and maybe put out of business) by the
various announcements. As the NY Times wryly put it, “How do you know if
you’ve created a really great, useful iPhone app? Apple tries to put
you out of business.” (The Times provides a handy list of apps now scrambling for a second act.)

In truth, quite a few of the things that Apple announced – such as a
basic to-do list app, and ways of storing web articles offline for later
reading – have become such fundamental needs for so many people that
they deserved to be part of the core OS. Unfortunately they are also the
bread and butter of many niches developers who saw the same need and
leapt to fill it in the intervening years. They will have to rethink and
improve what they do, and many of them will I’m sure.

Such is life in the shadow of an ecosystem behemoth. Apple giveth
(App Store to give independent developers more visibility and access)
and Apple taketh away (obviating the need for those apps in the first
place).

Apple has been pretty consistent in adopting good ideas from third
parties into its core offerings. Perhaps most famously, Apple introduced
the Dashboard feature (a precursor to the iconized app view on the
iPhone), to loud complaints of it ripping off a third party developer, Konfabulator who had created something very similar.

As problematic as this can be, it’s all part of Apple’s plan. Chetan Sharma put it succinctly: “Apple’s goal is to commoditize the software, Microsoft’s goal is to commoditize the hardware, Google – both”

Apple has high tolerance for making software free, even if it makes
life painful for its developers, because it makes almost all its profit
on hardware. For the time being at least, Apple has enough strength
and/or momentum relative to Google, Microsoft, media companies and
service providers that it can thrive with this approach.

The Mainstreaming of Cloud Services

The announcement of iCloud was met with both enthusiasm and incredulity.

Apple has been firing on all cylinders for years with hardware and
software, but has consistently stumbled with services, whether it be the
expensive and lackluster MobileMe (the launch of which even Jobs had to
admit at the keynote was “not our finest hour”), or the weak reception
to its music “social networking” service Ping. (This isn’t a new
phenomenon – anyone remember eWorld?) The only service area where Apple has really sung is with its retail stores.

With iCloud, Apple is cinching up the ecosystem it has painstakingly
built up, cinching it so tight that it will become increasingly
difficult for others – even ones as big as Google – to crack open.

MobileMe was an expensive, under-performing sideshow, but iCloud aims
to reach deep into all the other Apple devices and make them all work
together better. What was announced on Monday is surely only a hint of
what lies ahead in the next 18 months for iCloud, iOS, and OS X all
finally getting in sync.

Ironically, iCloud aims to improve on what was arguably the worst
part of MobileMe – iDisk, a basic cloud storage feature. Given Jobs’
obvious frustrations with MobileMe, I can’t believe he would let yet
another half-baked attempt out the door, especially not one that is now a
major strategic piece of the puzzle. Based on the massive data center Apple has invested in, they’re not joking around.

Since the iPad launched, its lack of a file system has meant it’s not
a true laptop replacement. One of the brilliant ideas about Dropbox is
that it essentially puts the file system in the cloud and moves it off
the device entirely. iCloud apparently opens the door for the same
thing, and with even superior integration. Today with near ubiquitous
broadband and 4G/LTE networks starting to roll out that offer home
broadband speeds while mobile, this suddenly becomes a workable
solution. (Bandwidth caps, tiered pricing, disappearance of
all-you-can-eat data plans? Yes, there are flies in the ointment, but
the longterm trend is clear.)

Linking Cloud, Apps, Devices, and OS’s

Consider two things that were discussed separately in the keynote:
journaling in the next rev of the OS, Lion (which means no more saving -
a file is continuously saved as it’s worked on), and continuous cloud
syncing. Voila – you have your most up-to-the-second work constantly
saved to the cloud, and made available on every other device.

My feeling is that iCloud will prove to be similar to IBM launching
its PC in 1983. Prior to that point, the PC market was highly fragmented
and dominated by niche players, and had little mainstream appeal. The
arrival of IBM on the scene gave PCs a stamp of credibility and
stability, and they gained sharply more acceptance. IBM made PC’s “easy”
to get into, made them relevant, and created the archetype which others
would mimic for decades.

Apple pulled off the same feat with mp3 players and smartphones, for
largely the same reasons. So it will be with iCloud. Cloud services are
not new (neither were mp3 players or smartphones), and the fact is that
much of our critical data already lives in the cloud, via various web
apps, service subscriptions, and email. But until now the various
services have been poorly integrated, and offered by startups that many
people don’t feel comfortable handing their data over to, whether for
security or long-term availability/stability reasons.

They haven’t been ready for the mainstream, and iCloud will come to
be seen as the turning point which changes that. For consumers who don’t
yet get the relevance of cloud, the media syncing across devices
provides the carrot to get into the concept.

MG Siegler looks at the different approaches to the cloud
being taken by Apple, Google and Amazon, and notes that “Apple’s belief
is clearly that users will not and should not care how the cloud
actually works.” Exactly. This is what Apple does best – take
complicated things that most people don’t care about, and makes them
easy and understandable for a mainstream audience.