S3
Cloud Based Service Applications - Google App Engine, Amazon Web Services, And More
Tuesday, April 8th, 2008 | Beta, Business, Marketing, New Beginnings, News, Tools, WebApps | No Comments
Hewlett-Packard just officially announced the release of HP Upline, an online storage service that serves as both a personal back-up source as well as a way to share files with friends. Not only that but it starts at only $4.99/month for unlimited data transfer & storage, making it ridiculously affordable compared to alternative large-scale backup methods. This announcement comes shortly after the successful launch of two similar services, Sugarsync and DropBox. The apparent need for a one-stop service provider that allows for data-backup, file-sharing, and cross-platform syncing has finally hit Enterprise level it seems. SugarSync service boasts the ability to synch not only your mac and pc, but also any mobile device that you may have as well. On top of that it would serve as a backup for your files should something happen to your equipment. The demand for the service was so high (their 45 day Free Trial may have something to do with it) that it actually crashed their service two days ago, resulting in a blog posting from their CEO Gibu Thomas explaining the situation and their progress towards fixing it. DropBox is currently still in a closed beta, but utilizes Amazon’s “cloud” based S3 service to do the same thing as what HP’s Upline is promoting - a virtualized synch/backup service that works on all platforms as well as can be used to share files with friends.
Speaking of Amazon, who so far has been the undeniable leader in Web Services, they too experienced a similar outage when their S3 service suddenly went down for a few hours a few weeks ago - leaving developers stranded as the entire purpose of using this service was to replace the need for any localized storage of data. The difference was that no one from Amazon has come forward to apologize or even acknowledge the error officially, the only mention was via a tech support representative after their forums were flooded with a battery of irate users reporting the error and asking why there wasn’t some sort of monitoring software/services in place (good question). Apparently Amazon either didn’t know about the outage, or simply didn’t send out some form of communication letting them know of the problem.
For a company to tout themselves as the replacement to purchasing/maintaining localized hardware to service enterprise level applications, this type of outage is simply unacceptable in my mind. Then the lack of response regarding the issue becomes an added slap in the face. Their complete lack of communication with the developers who relied on them was well documented across the web via the blogsphere but I sadly did not find it anywhere in the press. Why should it have been you might be asking? After all, it’s just some web-service. Their cloud services are currently being used to house large ammounts of critical data in applications that range from credit card processing portals to health-care based applications used in hospitals. Three hours of down-time with no explanation, warning, or even corporate outreach could have been catastrophic, even deadly. A great explanation of the incident can be found in a ZDNET article written by Michael Krisman where he even points out one worried developers plea, “But what do I tell my clients?” For a different perspective, I suggest you check-out what Bob Lozano had to say about the early-morning incident in his Appistry Blog (which also has a few other great articles on cloud/cluster based services if you’re interested in learning more about the service). S3 being down for any amount of time crippled a few companies who relied on it for data management…. and if you read the forums posts on Amazon’s site you can see the wide range of firms/industries who were helpless until the issue was resolved.
Since then there still hasn’t been any news from Amazon other than a small message explaining that the problem has been fixed. That doesn’t give me much confidence in the reliability of their other services, particularly their just recently announced launch of “Elastic IPs - Static IPs for Dynamic Cloud Computing,” This recent development would mean that you could not only use their virtualization/cloud services to store, manage, and transfer your database driven info but could also host the application that uses it through them as well. Throw in their Fulfillment Services and you have yourself a fully operational virtual company! An interesting thought was brought up on Brian “Krow” Aker’s blog (one of the smart ones who had a hand mysql + slashdot “back in the day”) comparing what I used to have running on a beat-up developement box in college and all these new offerings:
LAMP ->
Linux -> EC2
Mysql/Memcached -> S3/SimpleDB/Queue
Apache/Asterisk -> HTTP/Mechanical Turk
PHP/Perl/Python -> REST
Not to mention Google also has been aiming at getting in on a piece of the action. At last night’s Campfire One Google announced that they are now offering hosting for enterprise level web applications via a new tool called App Engine. Rumor is they also are partnering up with Salesforce (who have their own Appexchange service) in the upcoming week as well. Currently Google is giving away access/hosting for FREE to a limited 10,000 developers . I was lucky enough to snag an invitation, and although I have only had a few brief moments to tinker around with the site, it looks extremely promising at first glance. It currently only supports the Python language, however they are working on adding additional platforms soon-enough (i.e Ruby on Rails, or at least I hope seeing how it would be a perfect back-end for a cloud-based application of this magnitude).
According to their Google App Blog - the development environment provided is based off of how they themselves manage all the services/data, just rolled up into a more scalable (as in down-scale) commodity that they could then sell. This includes being powered by Bigtable and GFS - two services that have enabled Google’s search engines to run as fast as they do while parsing through petabytes of data to find you whatever it is you are searching for.
Soon the days of having to beg VC’s for startup capital to fund server farms will be replaced by virtualized data-flow…. I can tell you this much at least, all of these services give you the tools necessary to build whatever application you think might be the “Golden Ticket” - And It’d be a lot easier to get that $400 Million in funding by letting your potential investors actually use the product you want them to buy into. You then are selling them on the experience itself, leaving you free from having to dress in a suit to show off some slides of just how the experience “might look like.” And we all know that death by powerpoint is never a fun experience……
Search
Find A New Job:
Other Places You Can Find Me:
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
