Integrien blog | Integrien

A 100 Percent Virtual Environment Coming Soon…

As with any technology, there is bound to be a con with every pro. Take smart phones, for instance. With added functionality such as corporate e-mail and synchronization comes shorter battery lives.

Then there’s virtualization. Eric Siebert of SearchServerVirtualization says there are several arguments against implementing a 100 percent virtual environment. Among them:

• Single outages can cause many virtual machines to go down.
• The possibility of some software vendors not fully supporting your running their apps on virtual machines.

He goes on to say that it’s very possible to have an entire environment virtualized. But he advises against it.

“The cost savings that are typically seen by implementing virtualization will increase the more an environment is virtualized but you may want to stop at around 90 percent …” Siebert says.

We’re reminded of a recent interview with Diane Greene, the former CEO of VMware. In July, Greene shared her thoughts with Windows IT Pro – shortly before she was replaced by new executive Paul Maritz. When asked when she thought 100 percent virtualization would become a mainstream reality, she said anytime between two years and 10. (Talk about large windows.)

While it is hard to predict technology adoption rates, we think 100 percent virtualized environments will come sooner than 10 years. These environments, though, will require the best possible tools for performance management due to the complex interactions between virtual machines sharing a physical server and also spanning multiple physical servers supporting mission critical services.

Cloud Users Need More Than a Bill of Rights

Cloud computing has come under increasing scrutiny of late. From major glitches with Gmail and Google Docs to outages at Amazon’s S3 and Apple’s MobileMe, users are discovering that while a great concept, cloud-based services still have a long way to go.

Which is why we’re intrigued by the concept of a Cloud Computing Bill of Rights. Proposed by blogger James Urquhart, Article 1 starts out proposing “No vendor shall, in the course of its relationship with any customer, claim ownership of any data uploaded, created, generated, modified, hosted or in any other way associated with the customer's intellectual property, engineering effort or media creativity.”

While a bill of rights may appear grandiose, remember that all of a business’ applications and data can be located in the cloud. In sense, that IS the business. We think Urquhart has started a conversation worthy of future debate.

However, fundamental improvements need to be made before businesses go overboard with cloud computing. Troubles abound with cloud computing, including:

• An outage affects multiple users, which can have a chain effect.
• The Internet is not guaranteed to work.
• There are serious security concerns because crucial business data is stored remotely – and next to some other business’ data.
• What happens if a cloud company running an application decides to shut down?

Jack Schofield of the The Guardian newspaper in the UK opines on some of these problems, reminding readers in a recent story that Nikon is about to close its popular Fotoshare service and storage company Streamload has already called it quits.

A bill of rights is a good thing. Especially with articles guaranteeing that users are the ones who own the hosted data, not the cloud companies

But there needs to be better IT safeguards too. Putting your trust in someone else for all of your computing and data needs requires a serious amount of trust indeed.

Red Flags and Cloud Computing

Cloud computing, research firm Gartner declared in June, is an evolution comparable to that of e-business. While the concept of cloud computing is a great leap for business, getting a flexible and easy-to-use IT system in the cloud is also “inevitably harder” than the current hype has let on.

The latter declaration is also from Gartner. In the new report, “Hype Cycle for Emerging Technologies 2008,” Gartner doesn’t debunk hyped technologies such as cloud computing but simply tempers them a little.

“A lot of the hype will be around the promise that has been held out for a very simple-to-use, flexible approach to information technology,” says Gartner analyst Jackie Fenn. She adds: “It takes longer to get there than people tend to think.”

We agree with Gartner on both points. Cloud computing, which lets users focus more on applications and less on the back end, is headed to mainstream adoption within a year or two. Many small companies that adopt these remote services will likely enjoy significant savings on their capital costs. That’s why companies such as Google and Amazon have jumped into the fray.

However, companies need to think about issues such as analytics when signing on to a cloud service. In addition, they should learn how IT glitches in the cloud will be investigated; it may be impossible to do so in some circumstances.

There are other red flags, such as integration, that companies should be aware of before rushing to the clouds. As we’ve written about before, users of Google’s and Amazon’s cloud services are already too aware of these flags. 

A ‘Wake-up Call’ For Gmail and Apps users?

When Google speaks, everyone listens. When Google goes down, everyone talks.

That was definitely the case Aug. 11 when Google’s cloud-based Gmail went black, leaving many users without their normally steady service for two hours. The downtime affected not only 20 million-plus e-mail accounts – which include many businesses – but also Apps accounts. The blogosphere buzzed with complaints within minutes of the outage.

Google Apps gives everyone in an organization a custom e-mail address, word processing tools, a shared calendaring system and access to an intranet. It’s a substantial service, especially for the money. Apps and other cloud services, Scott Wilson points out in CIO Weblog, are “insanely cheap,” especially compared to hosting like services on-site. In fact, corporate Apps users at worse have to find $50 in their piggy banks per person per year. Fifty dollars! And Gmail is actually free for many corporate users.

Wilson goes on to say that Gmail and Apps are available 99.7 percent of the time. To get that level of stability, he says, you’d have to spend FAR more to run comparable applications in-house.

He’s right. However, millions of people aren’t relying on a normal company’s internal IT systems. With so many people depending on Google for their hosted services – their corporate livelihoods actually – two hours of downtime can lead to significant financial losses in productivity.

While impossible to gauge that loss in numbers – for Google or its users – Mashable blogger Mark Hopkins wraps it up nicely: “Given the amount of my digital life that is stored at Google, it’s a wake-up call.”

MobileMe: Not Apple’s Finest Hour

Since we’ve been on the topic of apocalypses lately, we might as well talk a little about Apple’s newest woes.

Just a few weeks after the company suffered glitches with its newest iPhone (dubbed the ipocalypse), Apple now is dealing with the firestorm surrounding MobileMe. The service, which syncs e-mails, contacts and calendars between PCs, Macs and iPhones, has been rife with serious bugs since launching in July.

Last week, an internal memo written by Apple CEO Steve Jobs was leaked into the blogosphere. In a rare mea culpa, Jobs said Apple needed to learn a lot more about Internet apps, and actually called the launch a mistake, since it occurred at the same time as the launch of iPhone 3G, iPhone 2.0 software, and the App Store. Said Jobs: “The launch of MobileMe was not our finest hour.”

MobileMe has been trashed for the past few weeks, even by Wall Street Journal great Walt Mossberg, who said he could not recommend the service due to the enormity of problems associated with it. Among other issues, he said the website was often “sluggish” and added that some calendar entries wouldn’t even load. That’s in addition to the launch bugs, which included servers that couldn’t keep up with traffic and e-mail outages.

In an Aug. 5 post in GigaOM, Om Malik wrote about the causes of MobileMe’s glitches. Om says Apple does not have a unified IT plan concerning applications. He also said Apple doesn’t even have a unified monitoring system.

We think Apple is one of the most creative and innovating companies around. But it’s lack of focus on infrastructure basics and IT operations management is disappointing to say the least.

Staving Off the ‘Virtualization Apocalypse’

Got to give credit for flair. At the most recent Black Hat conference, virtualization was the target of attack by Christopher Hoff, chief security architect at Unisys. In a speech titled “The Four Horsemen of the Virtualization Apocalypse,” Hoff said virtualization actually costs enterprise customers more, not less.

“Virtualization security,” he was quoted as saying by NetworkWorld, “can seriously impact performance, resilience and scalability.” He also said Cisco and other network companies are galloping off to adopt virtualization without fully understanding the security consequences of doing so. He called the situation “messy.” He went on to say capacity planning with a virtual network is going to get “difficult to predict.”

This isn’t the same virtualization line we’ve been hearing for so long. For years, we have been hearing that virtualization was the solution for all of our IT problems.

However, we agree with Hoff. Somewhat, that is. No company should rush out to adopt virtual networking without fully analyzing how to best configure such a system. And virtualization does add all sorts of challenges, from integration to security.

For those companies that do implement smart virtual networks, though, server virtualization management technologies can prevent the End of Days scenario Hoff paints.

In fact, implementing solutions that automatically learn the normal behavior of virtual components – so alerts are sent only in response to abnormal behavior – virtualization does save companies money. And when virtualized machines are optimized, they are easier to secure.

It’s wise to proceed into virtualization with caution. But with the right plan, and the right management and security solutions, you can stave off the bad guys -- Conquest, War, Famine and Death. And reap an impressive ROI in the process.

Bad IT and the Mercy Rule. When is Enough Enough?

In intramural and scholastic sports, the Mercy Rule is instigated when one team is so overwhelmed by its competition that further play would be pointless, embarrassing or both. Says Wikipedia: “In Little League baseball, rules usually call for the game to end if the winning team is ahead by 17 runs after five innings.”

In the world of taxpayer-funded IT projects, however, there is no such thing. Poorly planned and managed projects can go on. And on. And on.

According to a July 31 Congressional testimony by the Government Accountability Office, there are some 413 federal IT projects valued at $25.2 billion for fiscal 2008 that are poorly planned or badly managed. Troubles run the gamut from delays to poor performance.

We like context. So here’s some: Every minute of the day, the U.S. government spends nearly $48,000 on bad IT projects. Unfortunately, we’re not just talking bruised egos here.

In the hearing, the GAO said the Office of Management and Budget should provide more information on these problem projects. Detailed information was not made available, according to an article in Nextgov.com, due to fear of “embarrassing” the guilty. Aptly quoted in NextGov, Sen. Tom Carper of Delaware said they “need to be embarrassed.”

We know many federal IT projects are extremely complex and that mistakes are bound to happen, especially with projects without smart IT intelligence or predictive analytics. But things simply cannot continue this way.

IEEE Spectrum Online has been calling for an IT Mercy Rule for a while. We’d like to hear from you. Should there be such a rule – not to protect little egos but taxpayers? Or would an IT Mercy Rule only cost more, since bad projects would then have to be started from scratch?

-The Integrien Team

Psst: Amazon to Cut Down on Cloud Server Gossip

As we wrote about last week, Amazon’s Simple Storage Service was down for more than six hours July 20 – severely disrupting numerous companies that depend on its remote servers for computing and storage. Now, Amazon has let us all in on what actually went down that day.

According to the detailed report, error rates for the service began to “climb” at 8:40 AM. It would take until 4:58 PM for the service to return back to normal.

What happened? Amazon says “message corruption” caused problems in the communications between servers. (This, by the way, also happened about a month ago). And it took nearly two hours to figure out that the only solution to the most recent problem was to shut down all communications between Amazon S3 servers.

“We saw a large number of servers that were spending almost all of their time gossiping and a disproportionate amount of servers that had failed while gossiping,” Amazon says.

Why did it take so long to diagnose the problem? “We hadn’t encountered server-to-server communication issues of this scale before and, as a result, it took some time during the event to diagnose and recover from it,” Amazon continues.

We’d like to applaud Amazon’s honest report. We’d also like to applaud Amazon’s proposed fixes, including adding technology that reduces the amount of failed server gossips and instituting checksums to “proactively” detect corruption of system messages.

With these proactive solutions, as well as other application management tools (hopefully) down the road, Amazon won’t have to issue any more memos like this.