Enterprise Cloud Computing - Build Your Own With Cisco VFrame - Why Wait ?
21 August, 2008 by Greg Ferro Print Posting
I can see some value in external Cloud Computing, but why not just build your own with Cisco ? Take a bunch of leftover machines, that old storage system and get a demo version and make your own.
It seems clear to me that Cloud Computing is going to go down two quite different paths. The first is the path that Amazon / Google / Joyent represent. Enough said on those technologies.
But why not build a Cloud Infrastructure in your own Data Centre ?
The security and operational risk of handing your application out to something like Google means its unlikely to happen any time soon. Moreover, rewriting your business applications for their environments is probably never going to happen. No matter how I try, SAP on the Google is not going to happen in my lifetime.
The good news is, that Cisco already makes a tool that does, more or less, does a ‘do it yourself’ Cloud Computer. Lets look into that.
Cisco already makes a cloud solution
Sometimes there are so many products in the Cisco catalog that some of them get lost. But one product I have been researching over the last six months is Cisco VFrame.
Now VFrame is software toolset that automatically provisions VMware ESX servers, Cisco Catalyst switches, ACE Application Delivery Controllers1 , FWSM modules, Storage Arrays and Storage Switches and so on. In simple terms, it is programming environment that allows to ‘orchestrate’ the configuration of many separate technologies into a single process.
Let me ay that last bit again ” ‘orchestrate’ the configuration of many separate technologies into a single process.” Thats is the first part of Cloud Computing right there.
Looks Like a Cloud
Essentially a toolset that spawns a virtual machine, and all the supporting network infrastructure. VFrame will :
- automate the creation and deployment of VMware hypervisors, with all the standards and toolsets you specify
- automate the network operational changes including firewall, load balancing configuration.
- automate the creation and allocation of storage units to the virtual
If you ever created a system on Amazon EC2 (their so-called elastic compute cloud), you get the feeling that this is exactly what they have done.
Smells Like a Cloud
Lets look at the bit of Amazon EC2 :
To use Amazon EC2, you simply:
Create an Amazon Machine Image (AMI) containing your applications, libraries, data and associated configuration settings. Or use pre-configured, templated images to get up and running immediately.
Upload the AMI into Amazon S3. Amazon EC2 provides tools that make storing the AMI simple.
Amazon S3 provides a safe, reliable and fast repository to store your images.
Use Amazon EC2 web service to configure security and network access.
Choose which instance type(s) you want, then start, terminate, and monitor as many instances of your AMI as needed, using the web service APIs.
Determine whether you want to run in multiple locations, utilize static IP endpoints, or attach persistent block storage to your instances.
Pay only for the resources that you actually consume, like instance-hours or data transfer.
So if I have Cisco VFrame, a software tool that automates the following tasks:
- allocate and create a storage capacity and storage network from my existing storage system
- allocate and create CPU / RAM resources in the form of a VMWare ESX hypervisor
- creates the network modifications for firewalling and application acceleration
then that smells lot like Cloud Computing to me.
Feels like Cloud Computer.
So the VFrame software is able to connect to the Storage Array using various APIs, then create storage configuration and transfer the ESX image to the LUN.
The storage network is then configured to add the LUN’s to the MDS units.
The network is configured for IP addressing, switch ports, NIC teaming and any other switch configuration. Automated configuration for FWSM for security, and ACE for load balancing and application acceleration.
VI3 Remote Boot is used to start the ESX server.
Given that this all automated, we can use our standard build processes for every machine that is created, and the ongoing maintenance of these systems becomes simplified.
Elastic Computing
OK, where is the ‘elastic nature’ of the Cloud. The elasticity comes from the underlying nature of the technologies -
- VMotion takes care of scaling CPU and RAM resources from a possible pool of resources
- dynamic storage allocation on your MDS switch (combined with your Storage Arrays) provides scaling of storage capacity
- network virtualization allows you to move firewall and load balancing within the network to find bandwidth and processing resources.
Given that I can do this already, why am I outsourcing this to the Google or the Amazon.
Outsourcing usually fails
While I have seen outsourcing be successful, it almost never is. Outsourcing the computer functions of your IT may be an OK idea for mail other simple function like Office, it’s hard to perceive that outsourcing core compute functions will ever happen.
But we need something Cloud Computing in the change phase of our Data Centres. Funny thing, it’s already here.
Must be a Cloud!
I am sure I haven’t done this justice, and you should head over to Cisco’s Web site and get a look at the marketing material at www.cisco.com/go/vframe and get the full picture.
Here is thing, most people can build a test environment for a lot of this with the stuff you are throwing away. Get some of those old servers and storage arrays and build a test environment. Imagine how cool it would be to have your own “Cloud Computer”. Now that would be brilliant.
The thing I don’t understand, if why Cisco is not giving this more focus. They should have their entire “Social Marketing” machine out their pushing this like their lives depended on it.
There are some really amazing technologies in Cisco’s Data Centre strategy, and I reckon that a lot of people don’t know about them. This is just one of them. I have some more technical stuff coming in the next few on network infrastructure, so keep any eye.
Thanks for reading this far, feel free to comment and add information. Cisco Web site on Cisco VFrame is here.
Footnotes
- Application Delivery Controllers - yep, that the new term for a load balancer. Well, ok, they mimght do some application acceleration and app firewalling, but it is still a load balancer. [back]
Sounds Cool! The Data Center stuff is Insane. Maybe there will be a new CCIE Data Center. Seems like the trend. More to study for. My head hurts!
…or one could use Scalent V/OE, or Unisys uAdapt…
Check out slide #5 in Network World’s “10 Must-have Virtualization Tools”
http://www.networkworld.com/slideshows/2008/081808-ndc-virtualization-tools.html
(Yes, shameless plug!)
Yeah, but its in an article called virtualisation tools. All the buzz is about “cloud computing”, which is of course, the new marketing term for virtualisation 2.0.
I have never heard of Scalent so I guess you don’t have much of presence outside the US. Let me know when you get global.
No problem with plug, disclosure is the key to making it acceptable.
…shameless plug continued…
Yeah, we’re not yet quite the size of Cisco. Working on that.
But we’ve actually got quite a reasonable EMEA presence, direct and through various resellers — and recognition, see here:
http://awards.techworld.com/winners2008.asp?m_pid=0&m_nid=29322 (”Virtualisation Product of the Year”)
EMEA contact phone number on our site…
Greg,
I liked your post and wanted to let you know that I referenced you and your post in a Cloud Computing Examples post today. My thought is that a lot of the larger organizations with network and server support staffs will look at building their own Cloud Infrastructure as opposed to moving their data into the cloud.
Thanks for the post - Kevin Mullins
Greg,
Thanks for your insightful posting on building ad-hoc compute based clouds with Cisco’s VFrame Data Center offering. This product does have a significant development team behind it with good marketing material as you have read. While it may not appear that Cisco is heavily selling this product, Cisco is working with several large enterprise and service providers customers on the adoption of this technology.
As you speculated this is a non traditional product offering from Cisco and we are learning how to position and sell this to customers. There is a lot of device integration required when you think about tying the provisioning of servers, network and storage devices together as a set of malleable, virtualized cloud resources. Every week we add more devices to the VFrame DC compatability matriix. As you stated we encourage customers to try this product.
Thanks for visting our web site.
Bill Erdman, Senior Director, Cisco Data Center Technology Group.
Thanks Kevin.
Interesting article (and posts) about a topic that’s getting a lot of buzz in the market these days. I think Greg’s got it right as far as he goes, but there’s more to cloud than technology virtualization…one of the most interesting aspects of cloud is the fact that it also virtualizes geography. Not only can technologies be toggled on and off-line to meet business demands, but the underlying requirements for application performance (tied to latency and jitter), reliability (tied to physical redundancy), and data integrity (tied to data base replication) can also be achieved by linking multiple data center facilities over a shared large-scale network, all controlled through a common portal.
One final comment…the existence of a global outsourcing industry would seem to refute the comment that outsourcing generally doesn’t work. The only way to scale the technologies that Greg’s working with is through large infusions of capital which can either be provided by the enterprise or by a service provider who shares that investment across enterprises. There is an entire industry serving the needs of enterprises that is thriving between the DIY IT manager and the likes of Amazon/Google.
Great stuff!
The existence of a global outsourcing shows the limited creativity of senior business executives. Outsourcing was very fashionable and many companies now regret that decision. Increasingly the news tells of outsourcing contracts being reversed.
This is no accident. The business goals of an outsourcer are not aligned with customer and usually causes friction. Its an unhappy affair at best.
One day I might find a personal experience of a happy outsource, but it has not happened in the last seven years.
The external hosting cloud thing is temporary. The concept of Amazon/Google clouds will suit some people (mostly low cost startups making a lot of smoke but very little fire), but it isn’t going to change the world anytime soon. It’s hype today, some deployment and then….. just another IT choice.
Dear Cisco Marketing
See that, yep, that comment just above this, now _thats_ Social Networking.
Idiot comments on twitter, and poncy wiki’s don’t count diddly.
After Cisco’s announcement of the Nexus 1000V software switch at VMWorld yesterday, now does it make sense why VFrame was not front and center, until now? VFrame will be an integral part of the “automation” portion of Cisco’s Data Center 3.0 vision, which is coming to fruition with the Nexus 1000V, Nexus 5000, and VN-Link. Check out VMware’s announcements this week as well!
Since you speak like a marketing person, I will assume that you are. You haven’t disclosed any interest.
Sophistry on this scale must be addressed.
“now does it make sense why VFrame was not front and center,” No. In fact, I am willing to bet that the VFrame product management wasn’t holding back either and trying to sell their little hearts out to an audience that didn’t care. Cisco selling VFrame made no sense in the current market, after all Cisco didn’t “do” VMWare and customers didn’t believe it. VFrame hasn’t been “front and center” because it was unsellable. Thats why it wasn’t front and centre.
If VFrame is an integral part of DC3.0, it hasn’t been announced. Oh sure, I can see the vision, but no fire there. Smoke all you like though.
Nexus (in any form) is still 12 to 24 months from practical maturity. Measure my excitement in a years time.
Ouch.
1. I’m an engineer with a computer science background.
2. I’m 23 and relatively new to the field, but I am focused on the data center.
3. I loved this article…
Ouch.
then I apologize. I guess I have seen too much hype over time tend to be a bit cynical, perhaps a natural defense.
Hopefully some more cloudy thinking soon, appreciate your feedback
And you work for Cisco, you forgot to mention that.
My limited knowledge and experience prevents me from joining in the tech banter, though in due time, I will enjoy playing the devil’s advocate with such topics, not to instigate frustration or irritation, purely to advance my understanding in vastly conceptual matters.
I am currently doing a quasi RFP for a Networking class in which I selected Cloud Storage as my focus. All the recent hype of EC2, web 2.0, VDC-OS, and now the recent crash of the Stock Market potentially encouraging small, medium, and Enterprise businesses to utilize the Cloud Computing and Storage services to increase IRR and ROI. I find it amusing and a bit suspicious that the recent stock market crash parallels so timely with the influx of Cloud Computing.
Maybe it is due to my research; possibly it is Google Chrome and the simplified correlation that the public has raised awareness. Either way, from an Enterprise business perspective, I am grateful there are options such as Cisco’s VFrame Data Center and VMware’s VDC-OS.
Greg, I personally feel that Cloud computing and cloud storage via service providers can potentially be just as horrific as utilizing outsourced vendors in terms of the company data being in someone else’s hands.
Dennis, you nailed it on the head when you said that “the cloud virtualizes geography…underlying requirements for application performance (tied to latency and jitter), reliability (tied to physical redundancy), and data integrity (tied to data base replication) can also be achieved by linking multiple data center facilities over a shared large-scale network, all controlled through a common portal.” This is one of my key factors in selling my quasi-proposal.
Thanks for all your comments, and Greg, thanks for your article… I agree, but will someday disagree, and not just to ensure that you feel right.
Erin
Frankly, Service Provider or fully hosted Cloud Computing is about as likely and Britney Spears living a normal life. It seems possible, it should be possible but it won’t happen. This is the third outing of the Cloud Computing idea in the last 15 years and it still won’t be successful because the tools don’t exist that make it possible to identify and manage the blame points.
However, Enterprise Misty Computing is a definite, and all of Cisco’s recent moves are for that market. Of course, if a few service providers spend some bucks as well thats a bonus but it isn’t the endpoint.
Service Providers are big, slow, stupid and low profit. Enterprises are big, fast, clever and profitable. Which market would you attack ?