Despite the overwhelming marketing around clouds, the truth is the cloud ecosystem is still developing.  Standards and best practices are still developing.  While the cloud offers unprecedented flexibility, it also offers builders of clouds unprecedented ability to lock you in and hold your data hostage.  It also allows new attack vectors for malicious hackers.

Even as small companies have embraced the cloud revolution’s rapid infrastructure deployment, many businesses large and small have stood on the sidelines, wading into the shallow pool, but unable to swim to the deep end.  Larger companies are rightfully wary of trusting their critical data to a public or virtual private clouds.  Companies with established security practices use the public cloud for quick development or QA, but until a set of agreed upon standards exist to define what companies can and can’t do with your data, most companies won’t be able to leverage the public cloud or even semi-private clouds.  That’s why we created the Trusted Cloud Alliance, to establish those guiding principles and certify companies to a higher standard.

The current draft of our Guiding Principles are below.  We are working with our partners and major cloud providers to gather feedback and establish a final set of parameters the companies can use to easily verify the trustworthiness of a public cloud provider.

Guiding Principles:

• Cloud companies should establish a take over agreement with other cloud providers, should they run into financial trouble.  Current cloud providers represent a single point of failure that can be mitigated by establishing a clear line of succession and take over, should the run into trouble.  This will ensure that one company can not act as a domino and take down 100s or 1000s of other companies with it.
• Price changes are to be expected, but price changes should be fair, open and transparent, with adequate notice.  Companies should not change pricing without fair warning, in order to wring additional revenue out of customers, for already established features, while holding data hostage.
• Cloud providers should not randomly add hidden fees for cloud services.
• Cloud providers must have well defined ways for customers to easily move your data and applications of out the cloud.  Moving out of the cloud should be as easy as moving in.
• Dispute resolution should be clearly defined.
• SLA’s should be clearly defined.
• Cloud providers should define a clear and simple way for customers to get local backups of their data.
• Interface and User Interface (UI) changes can have repercussions for companies regarding training and useability. Cloud vendors should have old UI’s available for a set period of time before having to move to the new UI, and allow fair and adequate warning should a new UI be deployed.  UIs should not be changed over night with little or no warning.
• Planned downtime should require adequate notification and be at off hours to as to limit the impact to business.
• Change can take time, so cloud providers should display a clear Road Map to Trusted Cloud Alliance principles on their websites, with defined dates.
• Cloud providers should display a check list of the Trusted Cloud Alliance principles they have met or intend to comply with on their website at all times.

•   Blazing fast performance
•   Strong enterprise features
•   Reasonable & controlled costs
•   No single points of failure

Anyone who’s ever rolled out a virtualized infrastructure knows that getting all four of these things is nearly impossible.  Getting great performance is easy if money is no object.  Conversely, anybody can put together an environment that doesn’t scale.  This can lead to an environment which quickly falls apart as administrators add VM after VM and their storage array grinds to a halt.

The CTO of one of our favorite clients once said to us “you’re either a do it your self shop or a buy it off the shelf shop.”  90% of the time we think he is right.  However, we came to realize there’s a middle path, where you can get great performance at a reasonable price.  As we started to brainstorm our solution, we discovered that if we picked the right software and components, we could craft our micro-cloud to be something unique.

Let’s start with software.  There’s VMware and pretty much everyone else.  We want there to be a strong second virtualization vendor out there.  We really do.  But the truth is, there isn’t one.  Running anyone on Hyper-V is not realistic.  The worlds most unstable OS acting as the centerpiece of my environment doesn’t cut it.  KVM is a great hypervisor, but lacks many of the most basic features I take for granted with VMware, like live snapshots..  Citrix XenServer is a mess.  Xen itself is a hypervisor that was at best a kludge.  We think virtualization is moving towards commoditization, but it’s not there yet.

We strongly considered OpenStack, but it still lacks the features we wanted.  We want a hardened release, with a robust front-end and enterprise features.  In the future, we see ourselves building OpenStack clouds, but right now OpenStack is just not where it needs to be.  It’s a series of projects.  There are dozens of vendors building OpenStack products, but they all have an agenda that corrupts the basic tenants of OpenStack, which is to avoid vendor lock in.  When [INSERT ANY BIG HARDWARE VENDOR YOU LIKE HERE] wants you to deploy OpenStack on their hardware, it’s not because they are promoting open standards and choice.

In the end, we decided to go with vSphere Essentials Plus, which gives us patch management, a quick and dirty backup, live snaps, vMotion and HA.  This saves on costs and lets us spin up discreet micro-clouds that we can weave together with software, due to VMware’s excellent API.  VMware made the right moves with their API in vSphere 5.  The vCenter client uses the same API as third-party developers.  That guarantees that the API is clean and up-to-date.  There was no vendor choice here, but there was a way to save by educating ourselves on the correct licenses for our needs.

What choice should one make when deciding between a public cloud solution or building your own private cloud?  The fact is that public clouds are akin to leasing and private clouds are akin to buying.  When you lease, you end up paying the same price and you end up with nothing at the end.  Public cloud providers are all built on aging hypervisor tech, namely Xen.  They lack a lot of the features that we want.  There are also a ton of unsolved security issues.  How can you know your data is secure when it’s running side by side with thousands of others on the same multi-tenant environment?  Worse, what happens if you want your data out of the public cloud?  Will they export your VMs?  What happens if the public cloud providers go out of business?  Do you lose all of your data and infrastructure?  Now one might think that a company like Amazon would never go out of business, but much bigger companies have failed and will fail in the future.  Any company can collapse.  If they collapse with your data, it’s a disaster.  If they collapse with thousands of companies’ data, it’s an economic nightmare.   Private clouds are the only trustworthy solutions.

Next, we moved on to storage tiers.  This is where most of our cost comes in.  We built a small micro cloud for a security company that wanted to spin up VMs quickly for developers.  They purchased two EqualLogic SANs, three hypervisors and some Gig switches.  All in all their cost came to about $140,000 .  Not bad.  We like the EqualLogic units a lot.  However, this wasn’t going to work for us.  For starters, this environment was missing a number of our bullet points.  Performance is good, but not great.  With the EqualLogics you are married to only a few Gig links.  That’s fine for small deployments, but not for larger ones.  If you are going to use iSCSI you really want as many iSCSI lanes as you can throw at the system.  They also limit performance by forcing you to roll out the entire shelf as a single array.  This becomes a performance nightmare as disk I/O increases.  FC arrays don’t have this limitation.  FC was out for us though, due to cost.  So was NFS, because of speed and overhead.  It was block storage or nothing.  In addition, because people end up spending at least $40K on their SANs, they end up compromising on switching.  They end up with Dell or HP switches, which lack the robust feature sets, consistency and programmability of Cisco gear.  So in short, what did our client get for $140K?

• Inferior switches
• Not enough iSCSI lanes
• Disk arrays they can’t carve up

We can do better.  We realized if we could find a software SAN vendor, we could build our own storage tier and pack it with performance.  We settled on Open-E.  They have the majority of the features we wanted, like replication, iSCSI failover, and snapshots.  We also discovered that we could get some fantastic performance if we utilized the tiered storage concept.  For that we utilized LSI, who’s Cachecade cards let you build arrays of spindle drives and use blazing fast commodity SATA 3 SSDs as read write caches.  Now we could deliver some serious write through power.  Lastly, we combined our two U boxes with two quad port gig nics.  10Gbe is awesome, but again out of the question because it really ups the cost.  Instead we delivered 8 lanes of 1 Gbe iSCSI per storage node.  In total our SANs cost us roughly $10K a piece.  That is a lot less than $40K.  We picked up two for $20K, which gives us fail over and replication and helps to satisfy our no single points of failure.  So for $20K, we get the following:

• Fast tiered storage, utilizing SSDs as read/write cache
• Two SANs, with no single points of failure
• 8 Lanes of iSCSI instead of 1 or 2 with a low end vendor model

What about networking?  With the money saved, we were able to invest in quality networking gear, instead of knockoff switches.  We chose two 48 port Cisco 3750s.  There is a fantastic market for used Cisco gear.  These switches retail for $10K a piece.  We picked them up used for $3000.  Used gear is great because again, we are able to use two of everything.  We have no single points of failure.  The service contracts on them are calculated as a percentage of the retail price with depreciation.  No problem.  We’re not much higher than the cost of new Dell switches, but we end up with a strong command line, routable Vlans and standards that flow between network infrastructures.  We used the savings to purchase two Cisco ASA 5540s for about $2500 per firewall, used.   Now we have firewalls that can handle a tremendous amount of traffic and no single points of failure.

Last but not least we rolled our own servers.  When you buy HP machines, you get some great standardization.  Again, there is a robust used market for hardware like this.  We can utilize a generation prior, the G6 line, and fill the hypervisors with a few tiny SSDs in RAID1 for boot, dual 6 core Intel chips and 64 GBs of ram, all for about $5K.

This puts us at about $40,000 for our private clouds out the door.  Our performance and feature set will outperform most vendor’s off the shelf products.  Sure, we can’t compete against 10GBe Nics and EMC Clarions, but we don’t need to.  For a fraction of the cost, we deliver the following:

• High 9s of availability
• Breakthrough performance
• A robust feature set
• No single points of failure

At nearly 25% of the price for a similar product, what more can we ask for?

The only other question you should be asking is can your cloud provider do the same?

According to research conducted by MGI, McKinsey’s Business Technology Office, large data sets, known as big data, will emerge as the vital basis of competition establishing new surges in productivity, growth, innovation, and consumer surplus.

MGI conducted the study in five areas:  healthcare in the US, the European public sector, retail in the US, and manufacturing and personal-location data around the world.  The statistics are overwhelming.  For example, if the US healthcare system were to use big data in an ingenious and effective manner, the sector could establish more than $300 billion in value every year.  More importantly, two thirds of this would cut health care spending by 8 percent.

The study indicated that there are 5 obvious ways in which utilizing big data can constitute value.  First, big data can disclose substantial value by making information visible and available at a much higher rate.  Secondly, as organizations discover and save more transactional data in electronic form, they can compile more precise and accurate performance facts in regards to things such as product inventories or sick days.  This, the study states, will show variations and improve company performance.  Thirdly, big data allows a larger breakdown of customers. This results in the development of more accurately made custom made products or services.  Fourth, a comprehensive analytics system can considerably augment decision-making.  Finally, big data can be utilized to enhance the creation of the next set of products and services implemented by a company.

Statistics from the study indicated that there would not be enough talent available for organizations to benefit from big data.  By the year 2018, the US could be confronted with a shortage of 140k-190k people who do not have the skills necessary to benefit from big data as well as 1.5 million administrators who will be  needed to analyze the data to make important decisions.  Because access to data is critical, the proper protocol will need to be implemented to assure privacy, security, and liability in our big data world.   The study suggests that the positives from big data certainly outweigh the negatives.

The FAA has succumbed to requests from travelers to reexamine the law banning the use of electronic devices during takeoffs and landings.  Depending on the results of this analysis, we could receive an extra fifteen minutes to listen to music, chat, and surf the net at each end of one’s flight.  Would this change to more gadget time really be such a good thing, one might ponder?

Many feel this might be a potential step in the wrong direction, but also the reality of life in the 21^st century. Phones are not one of the gadgets that will be included in this reinvestigation.  How many people do you know or see who don’t listen to safety instructions or even bother to turn off their gadget when told to do so during a flight?  The argument for safety, however, during this time period is still a plausible one.

A group of high school students in my hometown of Chicago were recently put through a weekend experiment.  They went on a field trip as a group for three whole days without access to their phones or any other electronic gadgets.  When interviewed, many initially said that they felt lost without their phones and found themselves reaching into their empty purses and pockets.  Later they stated, that instead, they spent time conversing with friends, shooting hoops, painting, and exploring new things outdoors.  In my opinion, this was a good thing.  They didn’t think it was so bad either.

Perhaps the FAA needs to expand their law instead.  Maybe everyone in the US should be required to spend 15 more minutes a day interacting with each other in lieu of their electronic device.

The concept of “data mining” is quite popular with companies such as Facebook and Google who sift though vast quantities of data to discern what their users like and care about.  Microsoft recently created new software that uses this concept for one’s own personal accumulation of digital data.

This software called, Lifebrowser, compiles photos, emails, web browsing search history, calendar events, and other meaningful data from your computer, then carefully categorizes the significant events as they have happened.  The software possesses a “timeline interface” that allows one to browse through the timeline and contemplate memorable moments and events.   A volume control, allows the user to adjust how meaningful data has to be for it to show up on the timeline.  Another essential and useful feature is the search feature that enables the user to pull up events related to a specific topic.

Lifebrowser’s creator, Eric Horvitz, a prominent Microsoft scientist stated, “the motivation behind Lifebrowser is that we have too much stuff going on in our personal digital spheres.   We were interested in making local machines private data-mining centers that are very smart about you and your memory so that you can better navigate through that great amount of content.” Horvitz has 20 years of personal data stored in Lifebrowser.  Although it is still currently in its research phase, Horvitz is very hopeful that the public will get to take it for a test drive soon.

Charles reported that the 3CX system performed admirably by fending off the brute force attack.  The 3CX system immediately identified the invasion as a potentially dangerous entity, and therefore it was blacklisted.  After being blacklisted, the 3CX system was able to ignore all future requests by the attackers.  Although, it is probably a no-brainer to most,  Charles reiterated the importance of your firewall blocking unwanted traffic in the first place.

Are you the kind of person who doesn’t remember things even if you diligently make a list?  If you are human and sometimes lose the list you made to remember what you have to do, you will enjoy and benefit from the Microsoft One Note Application.  When utilizing this application, one can create endless notes and lists, save them, and take them to the grocery store or wherever you go with your smart phone.

This Microsoft OneNote application download is free and available for: Droid, iPad, iPhone, and comes pre-installed if you have a Windows phone.  The best features of this application are its large storage space and its user friendliness. Not only can you post your list on a website, you can make your notebook list available to friends and family so they can edit and work on the list with you.   Your kids can change broccoli to ice cream from their location as you wander the aisles of the grocery store.  Start creating, saving, and sharing your lists today by downloading this useful free application from Microsoft.

Google’s terminator style glasses project is set to launch and be available for purchase in stores by year’s end.

This smart phone alternative will be  android based, run on a 3g or 4g network,  and will stream information to your glasses in real-time.  Wondering what they will look like? Experts have suggested that the glasses will be similar in style to a pair of Oakley Thumps.  The glasses unique feature will be their navigation system. A scroll and click system which reviewers claim will be user friendly even for a technophobe. Steve Lee,  a Google engineer, who created the Google Mapping Software Latitude, is one of the instrumental people involved with the project.

Apple has taken a different approach to the same invention by focusing on straps to be worn on a person’s wrist. With a similar price point to smartphones of $250-$600, it will be interesting to see if it sells.  As a former Sony Walkman user, and current Ipod, Android fan, I can only get excited to think about what will be in stores next.

The primary issue I’ve seen so far with OpenStack has been the complete lack of enterprise builds, a la Red Hat’s RHEL version of Linux.  Enterprise companies can’t just roll out a whole rag-tag suite of different community projects that may or may not be integrated, upgradable and secure.  With so many projects now under the OpenStack banner, someone needs to do what Red Hat did with Linux, which is craft an enterprise version that is hardened, completely integrated and unified.  To do that, they will have to be selective and more conservative.  They will need to create a longer system lifecycle.  That means some projects will be left out in the cold, but that’s all right.  A smart company will create an upstream distro that is more cutting edge, a la Fedora and then a enterprise version.  I’ve haven’t seen that yet.

However, at least one company with some VC backed funding and some heavy hitters from NASA and the open source OpenStack project have created the first enterprise version of Cloud OS: Piston Cloud Computing.  Their version of OpenStack is called Piston Cloud OS or PentOS.  One many fronts they looked to have nailed it.  Check out the video below, where they rack four hypervisors/storage bricks and a 10Gbe switch, configure a single config file on a usb stick, plug it into the switch and then walk away while the entire rack configures itself.  That is some amazing stuff and we will be looking to partner with Piston in the coming weeks.  Their vision matches my vision of where cloud computing needs to go and in particular where OpenStack needs to go.  Public clouds are fine and dandy, but the real power is in private, on premise clouds built with OpenStack solutions.  Piston looks to have hit the nail on the head.

Piston Enterprise OS™ Demo from Piston Cloud Computing, Inc. on Vimeo.