As anyone new to Cisco can tell you, joining this company is no game of candy land. But I tell you, it is an exciting ride! We lovingly call the first couple of weeks for a new hire ‘the firehose’. Well, during the last three months that I’ve been at Cisco, I’ve been drinking from the firehose and focusing very heavily on the technical aspects of the UCS product line. Learning its nuts and bolts, kicking its tires, and even setting up a new system in our internal lab. Its been very exciting. I’ve been learning a ton, and can pretty much talk the value of UCS to anyone.
But this week when I was asked to give a general overview of Cisco’s vision of the Data Center lets just say I did a less than stellar job. When you’re down in the blissful technical trenches, its easy to talk all day about trade-offs of remote storage alternatives. But talking about it like you’re standing on the moon looking down at the earth is a different conversation all together, and one that I should have been more prepared to discuss.
I’ve thought about it some more and I’ve tried to break it down into something that is truly differentiating and doesn’t reek of generalities. To do that, I’ve had to frame it in the context of what is happening in the industry today. There are three parts:
1. What the future data center is doing
2. Why were not there
3. How Cisco will get us there.
The Future Data Center
Its about applications
The way we use technology to solve problems, to be more productive, and how to do it more efficiently is what IT has been about and what it will be about in the future. We do this today using fixed purpose applications: Microsoft Office Products, Email, Custom Apps that leverage the organizations data repositories, etc. The vision of most technology firms is that of a future even more focused on applications. Just ask Google, Apple, VMware, and Cisco.
In the future, we will still compose documents, but what will we use to do this? It used to be papyrus and lamb’s blood. Then we moved to ink, then the typewriter, and today in business it largely consists of Microsoft Word and some to Google Docs. Whatever our data center of the future looks like, it will be purposefully built to support our applications.
Applications are moving to the cloud or SaaS based
At VMworld 2011, VMware CEO Paul Maritz talked about how most applications today are being written by people younger than 35 and that they don’t feel the needs to adapt to the mold of software that has been written previously. Most of this software will be delivered using new frameworks, methodologies, etc. In essence: applications are being moved from being installed on your desktop, to being delivered over the network, or ‘the cloud’. Examples: Salesforce.com, Google Docs, 37 Signals apps, etc. Data is no longer stored locally on the end device. Its stored in a data center.
Why? To name a few:
- distribution cost = 0
- installation time = 0
- time to update the application on the users device = 0
- fraction of support costs, because lets face it: a lot of our support problems stem from users not being able to install the applications.
In essence: Once the SaaS app is created, it cost little to support. For the user: He can access it on any device, nearly any where, at any time. This is part of Cisco’s borderless networks vision. BYOD: Bring your own device. Where Cisco is focused on this is really making people feel secure about accessing sensitive data from anywhere on any device.
So what is the data center of the future doing? It’s hosting apps that can be run anywhere. If the apps can be run anywhere, then guess what? The apps can be hosted anywhere. The data center of the future is communicating with other data centers of the future. Its been called the Inter-Cloud. And much like Cisco paved the way for the Internet, Cisco will pave the way for the Inter-Cloud.
(Note: Do not confuse Inter-Cloud with private cloud. Inter-Cloud is like the Internet, but instead of many computers connected, there are many data centers connected. Those data centers will need to figure out how services like DNS, DHCP, etc will work across them. We’re still a ways off)
Why we’re not at the Future of the Data Center today
If your datacenter is doing nothing but hosting SaaS based apps most organizations will not be able to do it as cost effectively as cloud providers. (Amazon EC2, Microsoft Azure, Google App Engine, etc). This will only be more true as service providers continue to be more efficient. Some believe that Public Clouds will win the game.
I don’t think its that cut and dry. Certainly over the next 10 years I doubt larger organizations will be willing to give up their entire data center. So it will be a hybrid. And much like you can put solar panels on your house and give back to the power grid, you’ll be able to host workload on your own data center and seamlessly migrate it to the Public Cloud, or Inter-Cloud.
So what do we need to overcome to get us there?
- Security and Reliability
- Apps aren’t completely SaaS based.
- Still not as cost effective
- Transition costs
- Future protocols
Let’s talk about each point:
Security and Reliability
How many times this year have you read a headline about how some company’s data was compromised and now millions of people’s personal information is floating around in some cyber terrorists briefcase? I don’t think I need to dwell any more on that. Companies just don’t feel safe enough. Many individuals don’t feel safe enough. My wife won’t even use Facebook.
As for reliability, remember how EC2 had a major crash and caused outages in major web sites like FourSquare, Reddit, etc? What if 911 was running on it? I remember a story when some peers of mine were upgrading a network for the city. The customer asked them: “If we upgrade this now, are you sure we’ll still remain up?” The engineers were very sure. They’d done it before and had no issues. The customer then asked: “Would you be willing to bet people’s lives on it? Because if this goes down, 911 will be affected, and people can die”. The engineers decided to wait for a maintenance window. No matter how many 9’s of availability you have, you’ll always have a 1 with 0’s in front of it of non-availability.
In the future, if these security concerns can be overcome, we won’t care where the workloads are running, just as long as they know they are secure. Just like you don’t know exactly where your water comes from as long as you know its safe to drink.
Apps aren’t completely SaaS based
Think about how long it took from apps to migrate from the Main Frame to the client/server model today used by many desktop applications. Now forget about that number, because if you believe in the singularity, then the rate of change of technology is only increasing. For Apps to move from the client/server model of today to a SaaS based model it will still take some time. And if you’re not completely SaaS then you need to keep supporting desktops that have older clients on them.
This brings up an interesting side note about Virtual Desktop Infrastructure (VDI). In essence, if what I have written going forward is true, then VDI is just a temporary solution for many places. You won’t need a corporate approved desktop if you’re just running SaaS based apps. You just need something that boots up and has secure access. Today that is done with 0 clients, or thin clients. But guess what? Those thin clients will just be running apps. No need for cumbersome operating systems.
Personally, I can’t wait until I no longer need my Windows XP Virtual Machine to run just 2 or 3 apps. It is an immense source of pain to me. I can’t believe it took VMware this long to make a client for vCenter that didn’t require Windows!
Still not cost effective
Amazon and others tote the cost savings of outsourcing your data center. This is true for small and even some medium size businesses. But there comes a critical mass where outsourcing the data center, is less cost effective then having your own in house.
Zynga started out using EC2 but as the games got bigger, it was too expensive and so Zynga invented its own Z cloud. Today Zynga still uses EC2 for small tests or games that it is just launching, but for the real popular games, its just too much cost and they run on their own internal cloud.
Transition Costs
It takes too much work right now to move your app to the cloud. Where would you move it to? And if you moved it to Amazon, then isn’t that vendor lockin ? Where else could you go? The market just isn’t mature enough. So no thanks, we’ll keep our apps in house.
Protocols and Standards
What is the definite way to migrate a VM between Data Centers of different suppliers? What are the standard security settings needed on both ends to complete the handshake? How do data centers get certified that they can work together? Much of this still needs to be ironed out and I think you’ll find Cisco right there helping. The latest announcement at VMworld 2011 with VXLAN is a great start. (More on that in a second)
So even though public clouds aren’t extremely attractive today for many organizations, the concepts of a cloud are extremely attractive. Charge back, self provisioning, accounting can transform a data center into an extremely lean, efficient, transparent, and agile operation. That is why private clouds are being built. As private clouds mature, they’ll be able to share application hosting seamlessly through different data centers. That’s what the Inter-Cloud is all about and is what will give rise to our future public cloud model. This is the future of the data center.
The Road Map to the Data Center of The Future
How do we get to the Inter-Cloud? First start reaping the benefits of the private cloud. As more people do this, more technology will be developed, standards will arise, and we’ll be ready for the data center of the future and the Inter-Cloud. Cisco is already doing this.
At VMworld 2011, Cisco announced VXLAN which can enable virtual machines to migrate between physical servers in different subnets. Cisco has already been clarifying the gray area between the network administrator and the virtual server administrator with the Nexus 1000V. The only non-VMware dVS.
That’s all great stuff. But what should your organization be doing to future proof your data center?
Follow this roadmap:
- Begin SaaS based solutions transition
- Consolidate Hardware Infrastructure
- Virtualize infrastructure
- Build Private Cloud
- Hybrid Cloud
- The Inter-Cloud
This is how Cisco will get you to the Data Center of the Future.
1. Begin SaaS based solutions transition
Most organizations have already been doing this. Our applications are a combination of ones that must be installed on the desktop to ones that run straight from the Internet.
Universities like the University of Louisville and Case Western Reserve Institute completely outsource their mail to Google. The admins love it! No more headache’s with Exchange. No need to store all that email. The students love it too because its just like the GMail accounts they’re used to using.
Just like Google has solved the mail problem for hundreds of organizations, there other innovative companies have done this with other solutions like Payroll, accounting, CRM, etc. These applications will only mature and become more compelling.
Cisco itself offers a great SaaS based solution called WebEx for meetings and other collaborative products. Starting to use solutions like this is how organizations begin to transform from old world applications to the applications of the future.
Phase 1 summary: Get acquainted with and consider SaaS technologies: Try WebEx. Most likely, you’re already doing this.
2. Consolidate and simplify Hardware Infrastructure: Get control!
Internally as we look at our present day data center, it is siloed, static, and inefficient. There are all kinds of cables and networks intermingled causing massive overhead for IT. This infrastructure needs to be simplified and consolidated. According to a report last year, organizations spend 38% on IT personal. Thus, making these people more efficient is one of the fastest way to reduce operational costs.
Cisco offers its Unified Fabric solutions to help in this. It has two big impacts: First, it reduces management and cabling complexity for now you don’t need to buy as many devices. Second, it allows for multitenency. Fabric’s can be separated logically. Consider Cisco’s pioneering of technologies like VSANs where you essentially have multiple SAN devices running on a single physical SAN switch. This is a layer above zoning and once again collapses the need for multiple devices. With Unified Fabric, you only need to wire it once, then you can run whatever protocol you want over the network, whether it be Fibre Channel or Ethernet.
Unified Computing System or UCS is the other and important aspect of simplifying the management of hardware infrastructure. A single management tool, UCS manager allows you to do things you can’t do with any other server management platform. Administering legacy Blade architectures (anything non-UCS at this point), or rack mount servers is like writing assembly code to develop a web page. You may be able to do it, but using higher level programming languages like HTML and Javascript to do it is much more efficient. With UCS you’ll find something that used to take you 15 steps, now takes a single click. You’ll feel that you were writing assembly code. As an example, you’ll find you can update firmware on 120 servers with a single click. Think anyone else can do that today? Nope.
Phase 2 summary: Get your hardware under control. Consolidate with Unified Fabric & UCS
3. Virtualize the Infrastructure
The hardware capabilities of one physical server far excede the requirements of most applications. Indeed, they exceed the requirements of an entire operating system. Even my laptop is able to run 2 operating systems without even feeling a hiccup.
The next phase is to virtualize that infrastructure. With UCS, you’re most of the way there. But Cisco can make it even easier for you. We offer two major converged architecture solutions that enable you to have virtual machine appliances. They are called Vblocks and FlexPods.
With Vblock, Cisco has teamed up with VMware and EMC to create a virtual machine appliance. We’ve even created a company around it called VCE. VCE has its own engineers and support organization who create solutions based on this virtual machine appliance. They offer best recipes for all types of solutions. With VCE, you can order a single SKU, it comes to you preconfigured, and within a week you will be up and running. Try doing this with any other single vendor, or collection of vendors and it can take you nearly 100 days.
A Vblock gets you the best of breed architecture with the worlds leaders in storage, networking, and virtualization.
A FlexPod is a reference architecture with NetApp, Cisco, and VMware. NetApp is the fastest growing storage company and its innovative approach to storing data with its RAID-DP technology, De-duplication, and snapshotting make it a very attractive storage vendor of choice for many organizations. NetApp even offers a money back guarantee that if you don’t reduce your storage by 50% then they’ll give you storage for free. To date, they’ve never paid that out.
A FlexPod can come preconfigured with everything ready to go. Or you can buy the components separate, then register it with Cisco, NetApp, and VMware and we’ll tie the support on the back end.
Once you have started to virtualize your primary applications, its time to get serious and start virtualizing the serious stuff: databases and other critical apps. With VMware Fault Tolerance, your applications will be more reliable then they were on physical machines.
Phase 3 summary: Virtualize the infrastructure with Vblock or FlexPod
4. Build a Private Cloud
With your organization virtualized, you’ll already start seeing great advantages to managing your infrastructure. But it gets better. At this point the administrators are still provisioning machines. Its time to free them to move on to bigger and better things. Its time for the private cloud.
In this step, you automate the business processes for IT. You’ll build a catalog of services for organizations that can easily be provisioned and allocated on demand. This includes networking, virtual machine images, vApps, etc. You build capabilities to charge time to different departments for what they use. You start to bring incredible transparency to your IT organization. This is a good thing. Now people can see how valuable IT is and you may find IT is allocated more budget dollars.
In the last two years, Cisco made two incredible acquisitions. new Scale and Tidal. Since then, Cisco has adopted these companies in, combined their technology and created Cisco Intelligent Automation for Cloud.
With Cisco Intelligent Automation for Cloud, you get the only tier one server vendor who’s built a cloud solution from the ground up, and not just a rebrand of existing technologies. Its tested on the market and will only improve over time.
Phase 4 summary: Build the self service catalog with datacenter automation with Cisco Intelligent Automation for Cloud.
5. Hybrid Cloud
The term ‘Cloudbursting’ can refer to the situation when an organization receives spikes in demands for its cloud and must ‘burst’ the workload to a public cloud. This is the next step on the journey to the data center of the future. Your data center begins to establish relationships with other data centers.
Standards are still being developed to handle this stage. There are some interesting solutions in the open source field like Open Stack, and Cloud Stack that may aid with this journey.
Phase 5: Begin flirting with cloud service providers. Become a power user of Cisco Intelligent Automation for Cloud
6. The Inter-Cloud
Much like Cisco helped pioneer the Internet, and the Internet was built on top of Cisco networks, the Inter-Cloud will be built upon Cisco’s network. As we get to this exciting destination those who have taken the leap to the data center of the future will literally leave behind those who never started the journey.
“He who rejects change is the architect of decay. The only human institution which rejects progress is the cemetery. ” ~Harold Wilson
Phase 6: Become a part of the Inter-Cloud
Summary
I hope this article has been useful for any organization looking to understand the current data center landscape and see what an innovative company like Cisco is doing to aid people looking for solutions. In the end, building clouds and data centers is all about how we can help people work together more efficiently. This is exciting because as people are able to work better, more innovation will take place at an accelerated pace. Imagine as a result of technology the types of problems we’ll be able to solve both scientifically, biologically, politically, and organizationally. Cisco has a cool slogan about how “Together, we are the Human Network”. I think its safe to extend that to say: “Together we are the Cloud.”
If there is anything I missed here or corrections, please let me know. I’d love to get your thoughts.
Comments are closed.