Web 2.0 Watch

Web 3.0 Trends in 2010 - Did I say "3.0"?

I know, I know - there are plenty of people who haven't wrapped their heads around Web 2.0 yet so what's the rush right? Personally, I like to use the term, it is a useful handle.

Think of it as a further evolution of where we are presently. As Web 2.0 is the evolution of a PC-based world to a browser-based world, then Web 3.0 advocates assert that the next generation of the Internet will feature the semantic web (even more intelligent searches), more recommendations, more tailored data and delivered in real time.

I also believe that it means the evolution from the PC to more thin clients and mobile applications as a point of entry instead of personal computers and laptops.  Now that's a shift a long time in the making.

In that vein, Gege Gatt of ICON (a software application company based in Europe), asserts that several broad Web 3.0 trends are emerging this year.  Here's my take on it:

• Mobile Applications - Fusing mobile applications with locator software creates a lot of potential. One example would be loopt.com, which allows you to see where your mobile friends are on your iPhone. Another would be www.inMarkit.com, a digital bookmarking software that allows users to tag products and share them with friends on their iPhone, and ideally show the nearest retail outlets that carry the product.
• Interactive maps are a way to share information - we are all aware of Google Maps, but OpenStreetMap.org is taking things to the next level. Think of Google Maps meets Wikipedia. Open source mapping opens up a new useful way to look at data.
• Personal Organization - I consider this the continuation of a current trend. And an obvious one: as long as people are using smarter and smarter smartphones, the need to aggregate data from multiple streams (think Facebook, Outlook, and Twitter - all at once - I'd pay for that one) is paramount.
• Collaboration - Google Wave came and then went. (I'm not sure the strategy of dolling out invitations worked this time around.)  Companies like 280slides.net make collaboration on your favorite application, Powerpoint, about as easy as it gets.  Import your PowerPoint slides or create them in 280slides and you can get to your presentation anywhere in the world.

One other trend has been evolving for a while has been dubbed "Social Media Intermediaries" by ICON. These are add-on tools to social media products that make for a richer experience. TweetDeck would be an example as is Tweetag.org with the ability to detect trends in Tweets. Imagine a semantic real time search in the Twittersphere.

Gartner has weighed in on this trend by predicting that by 2012 over 50% of enterprises will use activity streams that include microblogging, but that internal Twitter use so to speak, will not gain much traction. This is a continuation of how Twitter is used today for outbound purposes not for internal communciations - its original reason for being. Salesforce sees potential in their version of Twitter for companies but I have to admit I'm baffled that people will use a new tool when there is a perfectly good one already establishedand available.Yet, I can see Twitter add-ons and mashups gaining traction over proprietary applications.

In the year to come, I'm looking forward to more richness of the current Web experience to be the guiding principal for many applications.

Ellison and the Cloud Rant - or was it?

When I was a college senior, I was at a career fair and at the Xerox booth, a recruiter was encouraging me to enter my information on their workstation. I said with all naivety, "Oh yeah, it's just like the Apple's we have at school".

Whoops. Mistake. Big mistake.

Xerox Guy was, to be kind, HIGHLY displeased. "No kid! It's NOT like the Apple! Apple is like OURS!" The rest of the rant...well, it was barely printable. Suffice to say that I got a rapid education in how Xerox pioneered a technology that Apple effectively marketed and expanded. Talking with others later (seems that I wasn't the first person to stumble into that bear trap), we were tempted to chalk it up to one part bitterness, and one part frustration that the pioneer gets the arrows, while the second wave of settlers makes the cash.

This fond memory came to mind when remembering one of the more interesting moments in the cloud world last year: Larry Ellison's rant again cloud computing. You can watch it here - I give this to Larry, he's entertaining whether you agree with him or not.

However, what Larry doesn't mention (why would he, he was on a roll) is that he did try a concept known as the "network computer" around a decade ago.

Here's how the NC2 was described back in 1999: "Ellison underlined the network computer's practicality, saying that upgrading the new machines will be as easy as swapping an Ella Fitzgerald CD with one by Miles Davis.

Unlike Wall Street, computing industry analysts voiced skepticism similar to concerns raised when the network computing idea was originally put forth. Kevin Restivo, an analyst at IDC Canada, predicted that Ellison will still have a tough time convincing many corporate customers that their users don't need a full-blown PC.

"He's trying to dispel the PC as a computing method, but the challenge is that there's still a level of functionality that people demand," he said.

Ironically, Ellison himself conceded that the network computer concept is somewhat overblown, as it will be used like such appliances as a television set, telephone or handheld to use the Web.

"Everybody gets so fascinated by this device," he said of the network computer. 'The interesting part is the Internet.' (Emphasis is mine.)

Sadly, the concept didn't work. Ellison lost $175 million over four years, and Oracle ended the grand experiment in 1999. But as Wired magazine said: "Just as [Ellison and Farzad Dibachi] predicted, a wave of cheap, underpowered computers with small hard drives and omnipresent Internet connections did challenge the traditional PC. Ellison’s vision of widget-based online software has become conventional wisdom. The network computer failed as a product and as a business, but it seeded an idea — and a group of technologists — that would go on to remake the computing world."

Of course this doesn't stop Oracle from working within the cloud in its own way be it with Amazon, or "grid computing", which is not considered "true" cloud computing, but for the life of me, I'm not sure why. Distributed computers acting in concert that could be geographically dispersed - hmm...

InfoWorld's David Linthicum felt that while Ellison's history lesson was factually correct, he misses the point on new cloud innovations such as IaaS and SaaS, which are great fits for the precursor thin client that Ellison attempted to sell 10 years ago.

Yes, "Cloud" is an inclusive term, but not just of technologies that existed, but of newer ones that have been developed. For example, Salesforce.com might have existed for a while before the cloud, but has laughed all the way to the bank as a SaaS, passing the $1 billion mark in revenue in 2009. Microsoft's Azure did not exist even six months ago as a IaaS (or even PaaS depending on how you view it), but it won't being going anywhere any time soon.

Larry, relax. You lost the battle, but your vision will win the war. Take a bow.

2010 is shaping up to be A Cloudy New Year...

According to the Gartner Group, 2010 is going to be the year of the Cloud. In 2010, 56.3 billion dollars were spent on cloud computing and virtualization.

The technology forecasting firm projects that by 2013, that revenue will nearly triple with $150 billion being spent on the general category "cloud services". As you can see in the chart below, the top strategic area in 2010 is projected to be cloud computing. It is instructive to note the differences in 2009 and 2010. 

The color codes are as follows: The light blue are technologies modified in 2010. The green areas are techs that are new to the Top 10. Note that Green IT, Web-Oriented Architecture and Mashups are all folded into Cloud Computing for 2010. And Virtualization, which Gartner slotted at Numero Uno last year is viewed now as Client Computing plus Data Center PLUS Availability.

If you see virtualization as one of the many services of cloud computing, then four of the top 10 slots are Cloud-related, and then you can add Social Computing (former buzzword: Social Software and Social Networking). Web 2.0 technologies are clearly dominating the 2010 forecast.

Ben Pring of Gartner has this to say:"Cloud computing is a broad and diverse phenomenon. Much of the growth represents a transfer of traditional IT services to the new cloud model, but there is also scope for creation of substantial new businesses and revenue streams” At this time, 83% of the overall market is focused on business processes delivered through the cloud.

When you consider the numbers that are being put up already, and taking into account the surveys that show that many companies (and governments) are still slow in adopting the cloud, there are fantastic growth opportunities. Example: A survey in the UK show that 81% of those surveyed recognized the value of cloud computing, but 75% of them wouldn't pull the trigger at this time due to the current economic climate.

On the other side of the globe, the analysis firm IDC stated "Future uptake of cloud computing looks strong. Over the next three years, as the use of cloud services expand from the domain of early adopters to that of the early majority, it becomes critical for IT vendors to develop strong cloud offerings, and play a leadership role in aligning their new cloud products and services with their organization, their traditional offerings, partner ecosystem, and customer and market requirements." (It should be noted that IDC had lower projected numbers for cloud growth, but foresaw large increases in cloud spending over the next few years.)

An exciting time in the Cloud world, no matter how you look at it. Services and tools, private and public are projected to explode. We'll be looking at it all year.

2009 Milestones in Cloud Computing

The last year of the Ought Decade (2000-09) was an eventful one for the cloud. Let's take a look at a few of the milestones:

• You had Amazon making headway, but also causing concerns with Amazon Web Services suffering EC2 with periodic outages, which was enough to make a CTO pause for consideration. After all, when any EC2 email address is automatically deemed spam, that isn't a very good sign that your house is in order.

• Major investment banks are looking to develop private clouds and incorporate SaaS, while the City of Los Angeles decided to take the plunge and go to the cloud with Google Apps, despite the misgivings of some. 

• At the government IT-Palooza that is FOSE, multiple governmental agencies are viewing the cloud with a mixture of hope and caution. The US government is the largest buyer technology in the world, to hear FOSE tell the story, yet Cloud Computing at the beginning of 2009 was considered a priority by less than half of those surveyed. However, the Federal Government has never been known as an early adopter - and I'm being kind. The new Federal CIO has vowed to change this mindset, and while attending FOSE, I saw many smaller federal agencies and individuals who were looking to the cloud provide inexpensive solutions to their issues.

• Cloud security emerged as the largest single concern to the proliferation of cloud solutions. A potential solution might have been found. If you are not familiar with the term "highly homomorphic encryption", you want to do so. Research by Stanford University's Craig Gentry is providing promise that data encrypted in the cloud will be secure. This solution would require enormous amounts of computing power, which is the specialty of the cloud.

• IBM entered the Cloud Wars (is that a fair term?) with LotusLive Engage, Connections, and iNotes. Meanwhile Microsoft unveiled Azure as a PaaS entry. This battle will only escalate in the years to come.

• I think it is safe to say that MySpace is so...2008. Facebook passed the 300 million mark in users; in other words the population of Planet Facebook is equal to the population of the United States of America - and growing, allowing Facebook to declare that they are now working in the black.When you see the integration of Facebook into various applications, both retail and business, it doesn't seem premature to declare a winner in this battle.

• Finally, did I mention that data security is a major concern? It bears repeating as 2009 featured the high-profile data wipeout at TMobile/Danger and their popular Sidekick cellphone. While Amazon had issues as well, the average consumer may not have been directly impacted, and world be hard-pressed to point to a inconvenience that they knew was cloud-related. But resetting your cell phone, only to find your Address Book and pictures gone - possibly forever? THAT strikes a chord. The lessons learned from this debacle will impact the perception of remote data storage for quite a while.

Who locks down your Cloud?

In a previous post, I wrote about the impact of cloud technologies in the financial services industry. The Royal Bank of Scotland is said to be moving to thin clients to better support SaaS applications, and UBS investment bank is making strides towards developing a secure hybrid cloud for part of their business. I've discussed earlier the potential advantages to such an approach. However, there is one area that is still a major concern to cloud clients.

Cloud security is the issue that concerns clients more than any one area. In fact, according to one recent survey, security concerns is the primary reason cited for hesitancy to move into the cloud. According to a LaunchPad Europe survey, nearly half of the companies surveyed cited security concerns as the reason why they would not adopt any cloud technologies in the next year. 

Additionally, Websense fired up their crystal ball, and predict that in 2010, Web 2.0 social networking sites will be special targets for attackers. Collaboration tools and Google Wave, Facebook, and Twitter especially would be special targets. This is rather natural on the face of it - any popular software that draws traffic will draw less than desirable attention. So what to do about it?

In a world where rule of law is inconsistent and hacking of financial sites and occasional attempted employee theft of information is a real threat – both UBS and Goldman-Sachs have had notable incidents recently - tight in-house control is required to keep business-critical data in-house. 

For example, in one case, a disgruntled employee attempted to sell part of the code base of business-critical software to competitors, only to be foiled by internal cyber-security (note: when downloading several gigabytes from a server, it will be noticed). One result of this is to allow only secure flash drives to be used by employees. On multiple occasions, there have been more conventional attacks from the outside. What complicates this issue is the understanding that what may be considered a felony in the US or the European Union may get a slap on the risk in other countries, which means that reward far outweighs risk. Remember the outcry over the reported widescale bootlegging of Windows and Office applications in China last decade?

There is hope on the security front though. The framework for this cloud security comes from the Sarbanes-Oxley regulations where disaster recovery and backups procedures are a matter of law. Any cloud performance and DR frameworks are guided by these principles. At this time, we don't have a firm definition of what a PCI level 1 compliance would be in the cloud - yet.

Some interesting points regarding this were raised by George Reese, founder of enStratus, a cloud management firm. In an interview on SearchCloudComputing.com, two comments really stood out to me regarding secure cloud development and the financial service case studies I mentioned. The first was this:

"I think people who are focused on 'security equals hardware assets I own or lease' are the walking dead. Owning your own assets doesn't make you secure..."

Note that the UBS idea of a private cloud goes beyond the owning of their own assets. Actually, their private cloud will allow a separate entity to own the physical assets, while the data stays in their control, and UBS owns data security.

Speaking of security ownership, let Mr. Reese elaborate:

"We've touched on the where the real security concern is, and it hasn't changed -- badly written applications and bad management policies. Those were the concerns before the cloud, and those are still the concerns in the cloud...[others] can't make you write a well-written Web application -- the only person who can do that is the developer. All hype aside...that's still what you have to be worried about."

Cloud security will require the same due diligence that a traditional setup would. I could not have stated it more simply.

An Azure-tinted Cloud at the Web 2.0 Expo

I had the pleasure of attending the Web 2.0 Expo in New York last week. It was a rewarding experience on multiple levels (more on that in a future column), but what I wanted to share today was a visit to the Microsoft booth and a look at their Azure platform.   

Azure is MS's new cloud development platform, best classified as PaaS (Platform as a Service) which allows developers to develop, stage, and deploy their applications in the cloud. Fully deployed, Azure removes the need for a datacenter. If you're interested in a deeper explanation of Azure take a look here to understand the underlying structure of Azure. A current user of Azure I talked to was the Social Media company Thuzi (pronounced Thu-Zee). Thuzi developed an application for Outback Steakhouse and their fans on Facebook. 

From the perspective of the user, you go on Facebook, become a fan of Outback, which allows you to get a redeemable certificate for a Bloomin’ Onion (speaking non-technically, not bad (smile)). Underlying, it's a bit more complex. As Thuzi developer Scott Densmore explains it, "Technically, we host the application through Facebook and iFrame. But we develop the application on Azure.  Microsoft takes care of the hardware, and the software that runs on the hardware. Being a Microsoft product, it supports .NET, and also Java, Eclipse, and PHP, among other technologies.”

Why not keep the hardware in-house? Densmore and colleague Mark Martorana: "We do social media, that's our business... and [we're] aren't interested in running a data center. Azure takes care of the staging where we do our work, and then when I'm ready we can upload in production. When were ready to upload into production, Azure can tear down the production environment, then port over the development from staging, then build it back up in production.” The ability to rapidly develop and deploy in the cloud is only part of the equation. The second part is dealing with steep spikes in demand.

You see in a week, Outback went from having 26,000 fans to over four hundred thousand. That’s 400,000 fans – a 15-fold increase in a week. In one day at the Expo, over 40,000 people became Facebook fans, a level of demand spike that will challenge system resources. Traditionally, one would buy a lot of new hardware to handle that load, a radical upscaling of the datacenter. Of course, the question that looms is just what in the heck do you do with all of that capacity when it is no longer needed – which could be as soon as two months later? 

This is where Azure does the heavy lifting of bringing up new instances when needed, and with their built in load balancer, Densmore says: “Azure figures out who is getting hammered and how to handle it. If we need more instances, Azure will spin up more instances as needed. When demand eases off, the number of instances will drop.” In any case, Thuzi doesn’t have to worry about hardware issues (acquisition and maintenance), which is a huge cost savings. Instead buying large amounts of storage and computing hardware that may not be needed a few months down the road, the cloud is great for covering variable demand. 

But how does one monitor performance? “We do our own utilization. We can add performance tools…”

You mean such as Keynote or other performance tools?

“If we wanted to, sure …Microsoft monitors capacity and determines the price. We keep up with CPU spikes and get an idea of how many instances we’d like MS to add…you see, we don’t want people going to Facebook and they say, ‘We went there, and the application doesn’t even work’, and then get all of that negative press. That would be bad for business.”

The other issue could be latency, the lag time in losing and restoring/replacing a node is not trivial. Densmore stated, “We can monitor at any interval. If an instance goes down, Azure will say, ‘This instance is non-responsive. You asked for 10 instances, but you now have nine, so let’s take him out and spin up a new instance.’ This can be done for a node, or whatever the problem is…”, and this action is done without a significant business lag.

Azure makes a convincing case as a option for a user looking at PaaS options such as EC2 and Google Apps. The ability to not only outsource the datacenter, but the software, while leaving performance monitoring flexibility in the hands of the user, makes Azure a player in a increasingly competitive PaaS and IaaS market. Thuzi would seem to agree.

Retailers Get Some Early Help Ahead of Black Friday

As we approach Thanksgiving week here in the U.S., much is riding on the outcome of the year's busiest shopping season. Retailers have been preparing all year and started holiday promotions early to get a jump start on holiday shopping. Focus has especially been paid to online shopping where convenience and predictions of special offers are forecast to send a record number of consumers - retailers will certainly need it.

One new source of customers for U.S. and UK retailers is going to be cross-border shoppers. Retailers in these 2 countries have been willing to deeply discount merchandise and U.S. retailers have the added advantage of a weaker dollar.  

So this week we announced 2 free offers to help retailers get ready and be successful throughout the 2009 holiday shopping season:

•  Free Monitoring for Retailers  - From now until the end of the year retailers can qualify for free monitoring of their Web sites and, for the most innovate retailers, for their mobile sites as well.  We'll handle all on-boarding and we're even making our performance experts available to help with troubleshooting.   This year, there's no reason to be surprised by a Web site outage or slowdown.

• Free Retailer Index - Over the years we've found that there's no better indicator for retail executives to know how well they are prepared for the holiday rush than our retail performance benchmarks.  These indices provide a clear measure of site responsiveness and stability in each retail segment.  This year we're offering free automatic alerts for 3 retail categories: apparel, books/music/video, and electronics.

Page B11, A Footnote to Internet History

You'd be forgiven if you missed the footnote published in yesterday's New York Times Business Day marking the end of AOL Time Warner.  On December 9, 2009, 9 years after AOL acquired Time Warner, it will be spun off to shareholders at a valuation of $3.5 billion.

When the merger was announced on January 10, 2000 the combined companies were valued at $350 billion.  Analysts and gurus of all stripes heralded it with much fanfare.  Here's a look back at what they said:

• "It is probably the most significant development in the Internet business world to date."

                    - Phil Leigh, Raymond James

• "Together, they represent an unprecedented powerhouse. If their mantra is content, this alliance is unbeatable. Now they have this great platform they can cross-fertilize with content and redistribute.” 

                    - Scott Ehrens, Bear Stearns

• "Today’s announcement really does change the tectonic plates in this world.”

                    - Christopher Dixon, PaineWebber

• "I don’t think this is too much to say this really is a historic merger; a time when we’ve transformed the landscape of media and the Internet.”

                    - Steve Case, former AOL chairman and chief executive officer

Much has transpired in the past decade and while AOL Time Warner could not live up to the hyperbole heaped upon it (which company can?) it has survived and at $3.5 billion is still valued higher than other heroes of the dot-com era, many of whom no longer exist.

While the AOL Time Warner merger was a reminder that all such deals are about salesmanship, the enduring AOL brand and traffic it garners (ranked #4 as of Sept. according to comScore) is a reminder that brands are built over years and can sustain all kinds of bad, discredited ideas.

AOL continues to be an innovative destination site relied on by many and continues to be a stalwart of its industry. (Not true for many of the institutions quoted above.)

Options, Derivatives, and Clouds - Oh my...

One industry that makes for interesting cloud projection is the financial services industry. Consider the amount of trading data that flies back and forth through cyberspace. Buys, sells, swaps, settlements, what-if scenarios and so on – millions of transactions a second – and that is just a small part of investment banking. With so much data comes of course a huge need for storage, backup, and disaster recovery (DR).

Recently, I had a very enlightening conversation with a few people, including an executive director (ED) at UBS in charge of supporting the Derivatives trading business, along with the middle and back office. It made for an interesting cloud case study if you will.  Here's the setup:

• Currently, the bank oversees a mixed environment supporting approximately 1400 servers. These multi-core servers (almost 900 are a dedicated internal cloud) are used for two things: 
• To store data and run the core business, which involves reporting and scenario development of equity trades. 
• Most of this operation isn’t in an organization-wide multipurpose cloud, it is on several dedicated sets of servers. But insiders see a real future for taking large parts of the current operation there.

This was a conventional data shop that originally resided in Zurich, Switzerland. In 2008, the operations were moved to New Jersey to be closer to the trading floor groups and their support teams. The next step of the evolution appears to be to a cloud environment that ensures data security. This will likely be a two-pronged approach:

• The first is how to replace the 1000's of mostly self-contained workstations that sit on the desks of employees. Answer: A staged transition to thin clients with the applications in the cloud. The advantages are costs are lower (far fewer licenses are needed), ease of upgrade (instead of support teams having to individually upgrade/service each desktop with its individual quirks, a simple upgrade can take place in the cloud at the source, then this is propagated to users quickly. There is precedent here: According to my contact, in the very near future, “…Royal Bank of Scotland is going 100% to thin clients with applications in the cloud.” 

UBS is not slated to go thin client on a corporate basis right away, but as various environments are due for major upgrades, they will be replaced by thin clients. All of the desktop applications go to the cloud. As a matter of fact, overall network performance is actually improved because there is an improved ease of load balancing.

Another “hidden” savings are the savings on UPS (Uninterrupted Power Supplies) units. A loss of power on a floor, or even the entire complex - is that crippling? Not really, as long as the server has a UPS there are no worries. Instead of restarting thousands of systems and support dealing with lost data from potentially thousands of users, the data can be recovered quickly from the cloud, in theory. And as an aside, the fewer power consuming appliances you have, the easier it is to develop a green environment.

• The second front will of course be with the data center and core business applications. Here is where things will be interesting. One possibility is a future cloud environment will feature a 3^rd party vendor setting up a private cloud – but one that is managed internally. Security for this set up will be two part: i) Human (UBS employees managing it), and ii) Multi-tiered in that security will be embedded in the API level. (Example: A token is required from anyone requesting a service, which will then be validated, plus authentication keys for users.)

I know what you're thinking: Can a scenario where the client is involved in the infrastructure be a "true" cloud infrastructure?

The answer is yes - in that the vendor will handle setup and any dynamic upgrades, maintenance and overhauls of the system. As far as a UBS is concerned, the hardware is invisible. As long as computing power and storage is there when needed and seamlessly, the rest is somewhat irrelevant.

In the words of my contact: “The third party will be responsible for all infrastructure issues, we give them our specifications for memory, and number of blade servers needed, etc…any hardware maintenance issues is for them to deal with. But UBS personnel will be the only ones handling the data and maintaining the software. When there are upgrades needed in the hardware, the vendor will be called in.”

From personal experience, I know that in a conventional environment, there is an expectation that you use the storage space allocated to you. Under-utilization is an inefficient use of resources, and almost as bad as not having enough space and having performance suffer. A flexible, but secure private cloud can potentially save millions of dollars hardware savings, infrastructure costs, reduced security issues, smoother performance and man-hours of maintenance time.

More later on the security aspects of a potential hybrid private cloud.

 

What Every Retailers Wants for Christmas and Why Many Won't Get It

Black Friday is less than a month away. In the run up to the end-of-year shopping frenzy we've just published an article on the state of online retail preparedness. You would think that all retailers have by now gotten their Web sites ready to handle the coming traffic deluge, right?  But many retailers don't know how their Web site is performing nor how it is experienced by their customers.

Ben Rushlo, a consultant who works with Keynote customers to help them get their Web sites in shape, puts it this way: “Some big players — I won’t mention any names — still have no external monitoring in place. They have a mentality that ‘we don’t need that piece,’ believing that internal monitoring and instrumentation can handle it and give them the data they need. But nothing can stand in for measuring your user experience from out in the Internet where customers live.” 

One of the companies that is doing it right is Sears and you can read an interview with their online executive on his approach for holiday preparedness and about new Web 2.0 features they are rolling out.