Peter Christy

Why Clouds Exist

At times it can be hard to understand what part of the Cloud discussion has substance and what is just marketing fluff. What I want to do here is argue that over the last ten years there as been a "perfect storm" of events that collectively led to today's Cloud world. Clouds are not just yesterday's time sharing or ASP offerings with fresh marketing lipstick. Here's the story.

Consider all that has happened over the period time since the Dot Com bust:

• The Intel architecture "won."  If you were planning a Dot Com venture your funders would probably tell you to use an Oracle database on Sun Solaris. In 2000, many businesses had strategic investments in the IBM Power Architecture (or HP's proprietary architectures). Today there is widespread agreement that Intel's X86 is good enough. Intel won the microprocessor battle. Without that architecture convergence there would be no Cloud.
• The network got a lot better. Broadband access wasn't common ten years ago. It is now. Without modern connectivity and bandwidth there would be no Cloud.
• Wireless networks expanded. Wireless build-out during the period was remarkable on a global basis bringing developing areas immediately into the networked world, bypassing entirely the wired phase.
• Mobile devices became powerful and smart. With Apple's iPhone deal with AT&T, mobile applications escaped the control of the mobile carriers. The evolution in wireless and mobile devices (and the rapid rise of consumerism -- employee chosen devices) has really changed the enterprise application and design model. This transition is an ideal time to also transition the front-end parts of the application to a Cloud implementation.
• We consolidated server applications into data centers. During the Dot Com boom, businesses spent profligately for IT, fearing that otherwise they would be "Amazoned." When the bust came there was broad retrenchment in IT spending (it's a cost center, remember). One of the easiest ways to reduce IT cost is to consolidate diverse applications into data centers where the administration and support can be local and shared. This cost-driven consolidation was an important precursor to today's Cloud world.
• Google demonstrated the promise of very large scale commodity computing. Ten years ago if you suggested building "mission critical" applications on anything other than "enterprise class" servers and storage you would be laughed out of the room. Google changed the game entirely. They showed that the most powerful systems in the world could be built from high-volume, commodity parts, and operated with high availability at remarkably low cost. In part the drive to the Cloud is motivated by the radically improved cost structure clearly visible downstream.
• Intel adopted multicore microprocessor design.  For years Intel converted semiconductor design and process improvements (the train under "Moore's Law") into faster CPU's by making the CPU's more and more complex.  In the end complexity yields diminishing returns. Early in the decade progress looked threatened: new processors got the same task done more quickly, but required more energy in the process. Finally Intel followed AMD and adopted a multicore strategy in which technology improvements are used to put more independent processors on the chip (more "cores") rather than to build more complex processors. It was a risky decision. It wasn't clear how many cores a desktop PC could use productively and the same problem existed in the data center where Windows servers were commonly built to each run on an separate server. The adoption of multicore designs enables a rapid increase the power of a data center of a given size and with a given power budget.
• VMware created commercial quality X86 virtualization and server consolidation occurred. Virtualization abstracts a workload from the specific details of the hardware on which it runs and enabled the shared use of a virtualized server by many X86 workloads with no requirement that they all ran the same "stack."

Virtualization is the magic lubrication that brings it all together: we consolidated applications into data centers; the victory of the X86 architecture and virtualization enabled running many disparate workloads on a shared server; multicore enabled an uptick in performance and power efficency improvements; Google proved that large scale data centers could be built from commodity technology and operated with amazing economics. The rapid evolution on mobile access and devices are transforming the requirements for enterprise applications providing a high-value motivation to change application front end architecture, just at the right time to leverage Cloud implementations.

There is hype and fluff in the Cloud discussion but underneath that there is a remarkable confluence of independent technology waves. In this case the substance is far greater than the hype.

Intel's AppUp: A Hybrid Cloud for SMB (a clever example of the power of virtualization!)

Like EMC’s VCube introduction last month (VCubes — a really clever idea) Intel’s introduction of AppUp -- an SMB offering leveraging virtualization -- is at once pretty obvious and quite profound. To understand this duality you first need to have a realistic picture of the discouraging nature of SMB IT today. For example, in an auto dealership you’ll probably find many different Windows applications each running on a separate PC or small server. If one of these servers fails or needs an upgrade it means ordering and configuring a new and matching server, carefully reinstalling the application and then grabbing the old data. The power of the server is rarely an issue. With today’s multi-core systems, a single small server typically has more than enough power to roll up and run all these applications on a single, shared machine. The question has always been how to into a more attractive and modern Cloud-like vision with as little disruption as possible.

Intel's AppUp is new and innovative model for delivery IT services to an SMB, developed and operated by Intel, and delivered to SMB customers by Intel business partners (including SP’s, MSP’s and ISV’s). Technically, the customer gets a small, virtualized server and runs the applications they choose (selected from an Intel “App Store”) as virtual machines on that server.  Software and updates are delivered over the network by Intel.  The user’s data resides locally on the server (providing robustness in the case of communications failures) and is backed up into the Cloud.  The overall system is maintained and supported by the business partner.

The AppUp business model is unquestionably modern. There are no up-front costs for the customer (the server is owned by the business partner). The customer pays for the apps on an ongoing, as used basis, based on the number of enabled users.

So on the one hand, AppUp is a straightforward usage of existing ideas, but on the other hand, it’s revolutionary.

• From the point of view of the customer, application selection and installation is dramatically simplified: click and go; and hardware maintenance is greatly simplified as well since the applications share a server, and since any dependence of applications on specific server hardware is eliminated.
• From the point of view of the SP, MSP or ISV who provides support to the customer, the task is greatly simplified. The hardware is generic, based on an Intel definition, and then implemented and sold by multiple hardware OEM’s.
• Usage is moved to an on-demand, pay-for-what you use model. All costs are fees for use rather than up-front capital outlays. There is no need to buy hardware or software up-front. In Intel’s business model, the business partner owns the hardware and lends it to the customer.
• From the point of view of the ISV it’s a way of leveraging many of the benefits associated with the Cloud (no setup, on-demand access, pay for what you use) without changing the software architecture.
• Since the applications execute on the customer’s premises, it avoids entirely the issues of having sensitive data in the Cloud as well as making the solution robust with respect to communications failures.

AppUp is another good example of how a relatively simple use of virtualization enables quite dramatic changes in IT.

Hyper-V: Is that the Fat Lady I hear?

For years I've waited patiently for Microsoft to catch up with VMware in enterprise virtualization, betting that Microsoft's experience with and investment in System Center would ultimately prevail and at least make the competition interesting. Last year things looked good for an intercept: Microsoft noted that more than 50% of Windows Servers attached to at least one System Center function, and that Windows Server was continuing to do well capturing more than 75% of enterprise sales. It certainly sounded like a winning combination, coming around the bend. But it's a year later and Microsoft still isn't really present in the Enterprise virtualization discussion, certainly not as a VMware peer. To be clear, Microsoft has lots of virtualization wins and System Center wins to be proud of, and Windows Server applications have lots of continuing success; my point is that they don't seem to have much traction being the overall virtualization manager. Furthermore, each year VMware integrates more and more of the system functions into the virtualization layer (e.g., components of the network, security). Over time these actions will be real competitive issues for the infrastructure incumbents like Cisco and Microsoft. (Microsoft was asked to comment on this but declined other than saying they disagree with the conclusion, citing IDC's reported 61% Hyper-V growth rate.)

Intel's Tiny Servers

Intel introduced a server CPU based on the Atom design (the low-end CPU that powers many netbooks). The motivation is interesting in this day of virtualization. With a virtualized CPU many workloads can share the processor resources and collectively get very high utilization, so why would someone prefer a much less powerful, non-virtualized processor? The answer is found in the large-scale, carefully engineered Web sites like perhaps Facebook. For these applications the smaller processor has significant benefits, especially if it is power-optimized and uses less energy to get the job done. These Web properties know exactly what the workload is and carefully manage the system to get full utilization out of the CPU, so the typical utilization benefit for virtualization isn't present. Also, the application designers say that a shared CPU is much more complicated in the case of some crash or failure of the CPU because many different workloads have to be repaired and restarted whereas with a single workload the error path is much more straightforward. The disadvantage of a non-virtualized design is that more explicit code adaptation has to be done to use a new and different CPU, but for now at least, clear proof that virtualization isn't a solution to all important problems.

More Virtualization Alliances

Microsoft and HP greeted the New Year by announcing an expanded cooperative relationship to accelerate the virtualization of key Microsoft Applications (e.g., Exchange and SQL Server) on HP servers (and increasing the attach rate for HP switches and routers too). Like the Cisco/EMC alliance, the basic intent is to make it easier to deploy and use virtualized systems. Whereas EMC and Cisco intend to incorporate applications into their consortium in the future, HP and Microsoft focus on the broadly used Microsoft applications Exchange and SQL Server right from the beginning. Microsoft's simple selfish goal is to accelerate the use of Hyper-V instead of VMware; HP's simple selfish goal is the grab more of the Exchange and SQL server share. As a relationship this is clearly in the "it's complicated" category. There is great value in Microsoft and HP jointly developing virtualized systems given HP's strong role in the server market, not to mention their EDS operations (a fertile field of interesting customers and potentially virtualized applications one might assume). On the other hands, HP and Microsoft both have many other partners bounding the value of exclusivity. And historically HP and Microsoft have been direct competitors in systems management, so we'll have to see how they sort out those pieces. In a discussion about UCS, a Cisco VP observed that you can think of modern virtualized systems as "mainframes" made from standard components (ultimately X86 processors, Ethernet connectivity and commodity disk drives). When IBM introduced the System/360 back in 1964 IBM made essentially everything that went in including the transistors, the disks and the software. Then we saw the "Cambrian Explosion" in computer diversity when companies like Sun emerged that didn't build any of the core technology other than the software, instead buying CPU's from Motorola, and disks and memory from other technology suppliers. Over time that diversity has thinned out by a process analogous to Darwinian "survival of the fittest" and we're now down to just three basic components in a standardized virtual fabric. Maybe now we're watching the early stages of the consolidation back to mainframe companies where the differentiation now comes at the virtualization layer and in systems management approach. Maybe these alliances are the reemergence of systems focus like in the age of mainframes. — pchristy@irg-intl.com

Virtualization and Security (Redoux)

Up until now, VMware treated security by saying as little as possible about it, knowing that arguing that virtualization didn't open up new security surface area just gave credibility to the thought that maybe it did (let sleeping dogs lie) which could only impede sales. Now that VMware is looking beyond the individual virtual machine, and thinking about the Cloud, they have started to talk (credibly) about how virtualization actually helps security. The issues are ultimately complicated (thank goodness there are bright analysts to think about them J ) but here's the basic story line. (1) If you want to talk seriously about the Cloud or even utility computing you need to talk about entire applications — virtual applications or Vapps — that include all the bits needed to get the job done, not about single virtual machines. (2) In order to deploy a Vapp to the Cloud (or even to a private utility or DR data center) you need to specify how it is to be protected. At first blush, a good answer is "just like you did before" except now the security can be largely provided by virtual appliances rather than the physical appliances used previously. (3)When specified on a Vapp by Vapp basis, and implemented by virtual appliances, the security can be as specific and customized as desired. For example, you can think about offering data leakage protection (DLP) tailored to the function and specifics of the application. You need to bind these security specifications in with the other application details so you know what else needs to be provisioned when the application is provisioned and you know where you can and cannot run the application (e.g., what other applications you can co-mingle with). Where it makes sense, part or all of the security can be specified parametrically (e.g., "PCI protection") so that as the regulations change, the specifics of the protection change automatically. VMW's observation is that binding security to the application and assuring it is provisioned as specified is a lot better than what we have now, where in most cases this is partially a manual process, and often requires coordination between multiple teams, all of which is complex, expensive, and prone to errors. The automation also makes it possible (and reasonable) to be much more specific about each application which is also good. Thus, virtualization makes security easier and better.

Virtualization at 50,000 feet

So here's a view on what's happening (the proverbial "big picture"): (1) MS has successfully queered the virtualization price structure making VMW offer the hypervisor for free. (2) That makes it clear that VMW isn't going to grow just selling the technology -- it's the higher level systems and management tools that you'll get paid for -- so they have repositioned themselves as a data center O/S company. (3) But that means that they are in a sense admitting that MS has been right all along -- virtualization is important but it's only one piece of a puzzle -- it makes VMW's position against MS scarier. (4) The vision is now clearer -- we want services that can be moved flexibly throughout a resource fabric (enterprise data center and cloud service provider) -- and we want an individual to be able to get to his personal (virtual) system from more or less anywhere and anything in the world. This is the attractive future that virtualization promises. The problem is that compared to virtualizing a single server this is really difficult system engineering and is going to take a lot of work to accomplish. Progress is going to start to slow down fairly dramatically as we collectively grind on these really hard problems. Don't get impatient. Sit down, watch a baseball game, smell the roses.

Networks Heat Up

A notable year in high-performance networking has turned a technical conference into a very relevant discussion of important commercial issues. Stanford's 16th Hot Interconnects Symposium (formally an IEEE meeting -- 27/8 August) has what looks like some remarkably interesting discussions including a keynote and two panels on Wall Street networking, a talk by Mendel Rosenblum on Virtualization and Data Center Networking, and a keynote and technical papers on how to deal with all the cores in tomorrow's multicore CPU's (see http://www.hoti.org/hoti16/program/ ). 2008 is the year when 10G networking started to be competitive, virtualization swept the data center, multicore impact started to be felt in earnest, and when the importance of computational intensive finance became all too clear. Even if the technology of on-chip optical interconnect isn't what lights your personal fire, you might very well find something important at this Stanford meeting.

Let the Hypervisor Wars Begin!

Microsoft seems to be feeling reasonably good about Hyper-V. Microsoft VP Bob Muglia reports it is functionally complete and performance complete and what remains is ringing out the bugs (as far as I can tell the stability is pretty good already). Mu said it was very surprising to him when they did their first performance testing against VMware ESX last fall and found the HV performance to be in the ballpark. When I first met Bob he was a Windows NT product manager. A key workload for NT was File Server and the competition was Drew Major's NetWare (as regular readers will now, Drew is in my pantheon of software gods). It took Microsoft until NT 4 to be performance comparable with NetWare so you can imagine Mu's surprise and pleasure to find HV Alpha code in the ballpark. The comparison between HV and ESX is complex. Microsoft knows it will take some time to catch up with the full VMW functionality. On the other hand, out of the box HV is supported by System Center, a much more comprehensive management system than virtual center. And just make things interesting and weird, MS announced and demonstrated System Center managing ESX, and MS support for management of key UNIX variants (along with OpenSourcing the integration technology). So this time MS is the heterogeneous, embracing vendor (that clearly tickles Mu too). How this all plays out is yet to be seen. MS points out that they know how to coordinate virtualization with Windows and key MS applications better than anyone and that should be quite valuable (e.g., to make virtual infrastructure memory management more efficient). In any case, MS is cranking up all the MS sales and competitive engines so things should get lively. When I pressed Mu at the MMS analyst meeting to say how we should observe competitive progress he said that after 18 months in the market if they weren't shipping more hypervisors than VMW he would be very disappointed. "Microsoft is very good at high-volume, low-cost software you know" (Hyper-V is priced at $28 if you missed that). Game very much about to be on.

System Center -- Front and Center

I've been attending the Microsoft Management Summit for three years now and the evolution is worth watching. The management effort began in earnest five years ago when Bob Muglia faced the fact that manageability was a liability for Windows server and hired Kirill Tatarinov from BMC to fix the problem. Kirill and his team developed a 10 year vision around the "dynamic systems initiative." The initial reasoning for the effort was that with good systems management Windows Server could have superior overall cost-of-ownership compared to alternatives like LINUX. The management systems team developed a pretty good plan and Muglia put his money where his mouth was. I think it's fair to say that five years into this ten year plan systems management is now a Windows Server asset rather than a liability. A little over a year ago the discussion started to move to focus on System Center (the Microsoft branding for the whole product line) as a profitable line of business rather than just a Windows Server asset. From what we heard at MMS we're estimating that the product line will close this year just south of $1B certainly a credible number even within MS. Perhaps more interesting, System Center (in turn built on SQL Server) is becoming a focal point for the virtualization strategy (virtualization management integrated with hardware, operating system and application management in contrast to VMware's "virtualization is everything" strategy) and also now at the core of Microsoft's Forefront Security as well. Muglia still describes it as five years into a ten year vision, and speaks to more sophisticated model-based applications (I'm expecting that some of these ideas will become clearer in the forthcoming Oslo release). The analyst meetings at MMS are getting a lot more popular (they used to be intimate little events).