A walk in the clouds
Many publications listing the top technology trends in 2009 have cloud computing on their list. It’s not something new, but as buzzwords go, this one is gathering more momentum. The reasons aren’t hard to seek - services in the cloud are becoming more mature, and bandwidth is becoming ubiquitous and inexpensive. Google, IBM, Cisco, Sun, Microsoft Salesforce.com have all done their bit. Many fledgling startups have thrown their hat in the ring. Grid computing aficionados are adding cloud computing to their interests. The acronym industry is in fifth gear with SaaS, PaaS, ITaaS, NaaS and all the other "aaS"es, though I find some definitions to be confusing at times. Our resident ’stir-the-pot’ blogger Doug Gourlay has chipped in with his thoughts and it seems that even some of the industry bigwigs agree to disagree on the definition.
The thing is, cloud services by whatever name are becoming popular, but they aren’t there yet, except in a few instances. And even it’s there 100% - ready - there will always be people, especially in branch offices, who’ll look for complementary offerings. Here are three reasons I think why:
• Trust - how much do you trust services in the cloud? And if you have the perception of security and trust, do you belong to an industry that mandates you NOT to trust? Regulations perhaps?
• Performance - When compared to the LAN, anything coming over a WAN link will have performance issues. Either because of the WAN link, or inspite of it
• Availability - Your cloud services may be available and always-on. But your access and connectivity to it may suffer, a corollary for the WAN link statement above. In that case, how survivable would your applications be?
I met up with my good friend Mike Wood, who’s the Director of Unified Communications and discussed a few of these ideas. He spent concentrated bursts late last year exploring business fundamentals of cloud computing, was brimming with new ideas, and as always, willing to share them.
As Mike points out, even branch offices that adopt cloud computing aggressively might benefit from having complementary “thin footprints†of their cloud-based applications available locally, to address any of the three issues highlighted above. That’s one of the directions we intend to take the Cisco ISR, AXP and other branch solutions within Cisco. A few flavours already exist, with Cisco SRST and AXP-based solutions.
I’m interested to know the cost benefits of this approach. It’s a key area where we have been focusing on internally, and have published short write-ups as part of the Cisco ‘Think Inside’ the Box Developer contest, and elsewhere.
We’ll approach this from a few different angles and see where it leads us…
Your thoughts?
Posted by Shashi Kiran at 08:34PM PST
Mark Hurtz Feb 2, 2009
In my humble opinion, grid computing and cloud computing are still over-hyped. I still don’t an ISP that will sign-off on the SLAs I’m requesting for our corporation. Grid computing is good for academics and theorists, but I don’t see practical applications coming into the mainstream in the near future.