What cloud services do consumers really want?  I think most consumers will, when they understand the concept, pay for cloud storage services.  They will (or do already) want a secure place to put their stuff, easy ways to put stuff in and get it out, and maybe a way to easily share stuff with others…stuff, of course, being content that they own.   Most consumers will, possibly without caring to understand how it works, be willing to use (useful) applications delivered via a software-as-a-service (SaaS) model (ideally for free).  Some will want to run special applications…like a blog.  Most of those would prefer to get access to these specialized apps for free as well.

Here, I just want to discuss cloud storage options for consumers.  In an upcoming post, I’ll discuss other service offerings.

[btw - I realize that "ownership" of some published content is up for debate...so, let's say, for now, that ownership is equivalent to holding a license to use, view, interact with, etc.  Consumers believe that when they pay a fee to gain access to content, that they should be able to conveniently utilize that content on any of the devices they own.  Maybe there are some concurrency issues to consider, and certainly, there are some sharing issues to figure out.  For example, if it's okay for me to loan you a CD, DVD, or BluRay disc (it is, isn't it?), then why can't I let you access the content that I put up in the cloud?]

Well…regardless of whether one should be able to do this, it will happen, is happening.  The question is how to deliver cloud storage services to consumers.

There are two main models that are being developed/offered:

  • A centralized model, being developed by Google, Microsoft, Verizon, Amazon, and others.  In this model, all data is pushed up to a massive storage infrastructure that is managed by the service provider.  As long as the consumer has internet connectivity, they have access to their content via a variety of client applications.
  • A widely distributed model, being developed by Cloud Engines and others.  In this model, the consumer purchases an interface device and a storage device (or an all-in-one NAS device), and operates his/her own storage infrastructure.  The service provider facilitates remote access from a variety of client applications and devices.

So, which of these models is right for consumers?  Let’s look at each from a consumer viewpoint.  Which is best for the consumer?  Which will win out in the end?

Centralized Cloud Storage Model

There’s a big push to build out and sell centralized storage services.  The marketing spin is all about ubiquitous access and data security, but the real goal is user retention.  Carriers need something, anything, to sell to consumers to lock them in.  Without this up-sell/lock-in approach, carriers would be vulnerable to consumer price shopping  and downward pressure on bandwidth pricing…the dreaded dumb pipe syndrome.  Ad-based services (e.g. Google) need to keep consumers coming back.  The more a consumer accesses a search provider’s services, the more the provider can learn about the consumer, and the more valuable ad impressions become.

How does it work?  It’s pretty simple.  For some subscription fee, or perhaps even for free, the consumer is allocated an amount of storage within the provider’s cloud storage infrastructure.  Consumers push content up to their drive in the sky, and then can access their files via client applications supplied by the storage provider.  The provider takes care of backups and availability (of storage, not necessarily network access).

There’s plenty of incentive for certain cloud storage providers to move consumers to a centralized model.

In this model, the provider charges consumers a subscription fee for some potential amount of storage, not necessarily for what is actually used.  Further, the provider calculates space consumed using the uncompressed size of files uploaded.  Of course, they compress files for storage (and possibly for network transmission), and while some files don’t compress much, some do.  The difference between perceived reserved space and space actually used is known as breakage.  The fees consumers pay for breakage are pure profit to the provider.  It’s like buying a gift card that is never used, or subscribing to Netflix and never putting any movies in your queue, or like joining a gym and never using the facilities.

While some services provide a small amount of storage for free (e.g. Google provides 1GB of free storage with Google Docs), this limited amount is almost useless for media storage…and, (I contend) most consumers will want to store their media collection in the cloud (or somewhere that is accessible from anywhere).  At 4GB each, a collection of BluRay movies will consume a lot of space quickly.  Allocations of 200GB and more will be common and, at this point, price becomes an issue.  Google charges $50/year for 200GB of space, and $100/year for 400GB.  By contrast, one can currently purchase internal hard drive space for just over $0.05/GB (retail).

Distributed Cloud Storage Model

A few vendors are now selling network attached storage (NAS) devices that allow consumers to provide their own “cloud” storage.  Still others are selling small network devices, with NAS software embedded, to which consumers can attach their own external disk drive via a USB interface.  All of these become, essentially, always-on computers residing securely within the consumers home network environment.  Some, provide remote access via services from the device vendor.

I think the folks at Cloud Engines (Pogoplug) are on the right track…not there yet, but close.  (Personally, I think they should have stuck with the plug computer form factor that they used for their prototype.  But, what do I know.  Maybe everyone else likes the oddball packaging with magenta trim.  At least they put black trim on the business version.)

For $99 you get the NAS interface device, free remote access service, and applications that provide transparent access from PC’s and mobile devices.  One major missing feature is support for DLNA and UPnP.  Any NAS device targeted at home users must support these protocols.  Consumers may not ask for it now, but they will…and soon.

A more complete solution is being offered by iOmega…Their StorCenter products offer RAID 1 (mirroring) support, and come with included backup software, remote access, and DLNA/UPnP support.  The remote access service, which sounds like may just be a dynamic DNS service, is free for the first year and $10/year thereafter.  If this really is just a dynamic DNS service, it should be free for the life of the product (imho).

For those consumers that are capable of rolling their own solution, the folks at CodeLathe present a very good solution with Tonido.  Their TonidoPlug allows consumers to create their own personal cloud.  It’s based on the (now) popular SheevaPlug platform from Marvell (the Pogoplug is based on the same platform).  The plug is similar to the Pogoplug, but includes DLNA/UPnP support, and CodeLathe provides many more client applications.

Another SheevaPlug-based system is the hybrid CTERA CloudPlug device.  This system blends the Pogoplug approach with backup to a centralized cloud storage provider.  The system, targeted at SOHO users, supports UPnP and Bonjour, which is Apple’s service discovery protocol, but apparently does not support DLNA.  It also provides sync to PC and sync to cloud features that may improve access performance for business users.  (This appears to be a simple way of providing offline access to files on the PC, and essentially auto-backup to the cloud.)  I think that keeping three copies of files in sync provides a lot of opportunity for error, and may be a point of confusion to some users.  I haven’t had any hands-on time with any of these devices, but this one seems to be over engineered and not simple enough for the ordinary consumer.  Again, it’s targeted at SOHO users, but may be missing an opportunity.  If it can be delivered affordably, this hybrid approach may be the right one.

Consumer Cloud Storage Requirements

So just what are the requirements cloud storage vendors should be addressing?

  1. Simple, transparent – the system/service must be easy to deploy and use…essentially invisible.  Let’s look closer at this below.
  2. Affordable – that is…cheap…say not more than $10/month for unlimited space (let’s assume 1TB is sufficient to be considered unlimited).
  3. Fast – at least it should be perceived as being fast.  This simply means that latency should be un-noticeable.  Media files must start playing quickly and stream at a rate that will support real-time playback of the content.  Other files would ideally open as if stored locally.
  4. Accessible – must be easily accessible from all devices the consumer may want to use to access their content.  (Again, let’s look closer below.)  Sharing features should allow the consumer to provide access to friends, ideally without requiring software to be installed on the friend’s access device.
  5. Available – stored content must always be available when the consumer wants/needs it.
  6. Reliable – stored content should not be at risk of loss.  This implies a need for RAID or some kind of integrated backup.
  7. Open – consumers should not be locked in to a solution.  They should be able to easily move their content to another solution.

To drill down a bit on some of these…What does it mean to be simple, transparent, and accessible?  I think it means that the solution must integrate easily with the collection of devices and applications that will be used to store and access content.  So, support for auto-discovery protocols like UPnP, DLNA, and Bonjour are needed to facilitate installation.  Support for common file sharing protocol (CIFS, for example) are needed to facilitate storage and access.  Native or embedded applications should provide intuitive access via mobile devices.  One should not have to download a file to open it in an editor.  The storage volume should appear to be a local device and the solution should facilitate multi-user access (i.e. file locking and perhaps block-level access).

Quick Analysis

So which of the solutions mentioned above meet all of these requirements?  The short answer is…none of them.  I think the iOmega StorCenter solution comes close, but appears weak in it’s remote access features.  Pogoplug comes up short in the reliability area.  I think the CTERA solution is too complex and perhaps too expensive.  And, the Tonido solution lacks RAID or backup and just seems too cumbersome for ordinary consumers.  The centralized storage solutions fall down in regards to affordability, accessibility and openness.  I really think they can benefit from a hybrid solution like CTERA’s, with modifications, but also need to find a way to deliver LOTS of storage for a low price.

Again, I must admit that I have not had any hands-on time with any of these solutions.  We’ll take a closer look in future posts.