Disrup the Disruptor


Platform

Platform (Photo credit: Geir Halvorsen)

In one of the recent interview the CEO from Pivotal, Paul Maritz, stated that he wants to disrupt the disrupter. The AWS platform offering is addressing a majority of the public cloud demands today and Pivotal wants to play here strong too.

So what is the disruption AWS is addressing?
“Fast”, DataCenter by credit card, to name one. “Easy” maybe another. “Cheap” is also stated. All this has nothing to do with a conventional DC, more the opposite, slow, complex and expensive is more what CIO´s today see in their environments. This is the disruption, there is only one missing piece, “Trust” or “Risk”. Is this not what DC are build for?
Let´s circle back in time, where the Mainframe was born, lets call that a platform one. The majority of applications written there have been determined the production systems. Of course there have been litte users and litte amount of demands, but the origin was done thru the developers.
History has shown that there was only one player left after 30 years. The programs still running and working, but hard to maintain and not able to be inline with current standards of customer experience.
With the age of the PC and more with appearance of Linux and x86 architecture the second platform was born, we called that client-server architecture. Also here the developers led the way with creativity and innovation. Many languages where born, ended up in the internet itself. The HW was still very expensive and followed moores law for years.  Microsoft got the dominant player. This generated the complexity in the DC, since the legacy has to work with the new stuff. Silos where build and middleware controlled the systems.
“Suddenly” Google and Facebook like companies where borne and storage, CPU and Network more or less have been given for free. They connected billions of users with millions of applications. The App century where born, lets call it third platform. And again the developers paved the way. From my perspective most of the current App`s are specialized browsers which have a nice interface to Informations.
Here we go thats the buzzword we will see more and more in the future. Information, which is data with meaning, was always generated in in companies and kept very close. Now OpenData appeared, mobile devices generate more data and Information suddenly could be enriched to get new business models in place.
Have, in platform 1 and 2, business leaders invest in DataCenters, not knowing how large the business will go, now they can purchase non the trip and on demand via a credit card, nice.
Many apps and services consumers today buy will leverage such infrastructure.
Hence, guess what, what moste of the the startups leverage this model.

 

Pivotal Labs

Pivotal Labs (Photo credit: teamstickergiant)

 

 

Now how to disrupt this growing business?
The answer is in the definition what business. Large global companies will have big troubles to run their business under the current offerings. Even the current cloud provider adopt fast and be very creative, there will be a lot of legacy which cannot transported. Like the mainframes not to the client-server architectures. The key is in combination, or bridging. Investments which have been taken and will be taken to support the current business models have to be adjusted for the next generation architecture. EMC`s federation approach is aiming for that. Complexity of Infrastructure will be solved by Converging and Software Defined – X concepts, Optimization and orchestration of the infrastructure vmware is leading the industry by a magnitude and pivotal will provide the open platform architecture to combine the business needs form the current and future demands.
When we talk technology, Cloud Foundry and major parts of Pivotal One, which comprises many supreme technologies are open and crowed developed to exponential capture the great ideas from this planet. This is the real disruption. In Platform one it was very country centric, Platform two was dominated by the thinking of the western world, the next platform has to address a global demand and population.

Enhanced by Zemanta

Why Standards matters


Do we really need standards in infrastructure?

Mandelbrot, Designed by Frax for iPAD

Mandelbrot by Frax for iPad

Since a couple of years I have discussions with CIO`and other technology employees in global companies around standards. In fact the introduction of SAN as a protocol standard enabled the largest consolidation and optimization in the DC, started early 2000 until now. One of the reasons is that there has to be a maturity of the market and also a common sense that there is no benefit of defining the “company” standard.
As we have seen in the telco industry, over years companies have define there own phone infrastructure, it has total changed now to a VOiP driven model with a lot of benefits. Same we see nowadays in the DC with the appearance of the converged infrastructure. Hence this approach is not yet mature enough to have one single standard and the manufacturer of infrastructure and VAR´s define there own way. This currently has major impacts at customer sites since there will be a next generation converged model with is more defined thru organizations like SNIA and open stack models, or it will be a 3rd platform approach more likely, which may be called cloud.
What ever the next years will bring the key here is that decisions are not driven only on price more on the TCO and the future of the proposed architecture.
It is obvious to see that innovations cycles will increase and the demand for more flexibility on the business side with drive different infrastructure needs.

What happened on the manufacturer side?

Keep it simple is the demand which was given by the creators of the 3rd platform. However, in the enterprise of global customers the IT still runs in the 1st or mostly on the 2nd platform. This has to be taken in considerations. Associations like SNIA take this as basis and define standards which span the bridge between the new innovative and the current business demand IT requirements. Industry standards take a while to establish and of the manufacturers develop there own “standard” to keep customers closer. In the early stages of new technology this many be very convenient, but with maturity of the approach the move to an industry adoption will become necessary.
The same happened currently in SNIA. The standards around the traditional SAN is defined and also management capabilities like SMI-S are defined and adopted. New areas like Big Data, Object Storage, Analytics and Flash are more in the definition phase and the manufacturer define there own strategy and API´s. SNIA is working deep with the vendors to get the industry better shaped here. Seeing still a lot of startups appearing and innovation happening there too, the industry standard definition is still in the starting points, but customers would be wise to ask for certificated products to make the TCO and technology adoption more efficient to serve not only the cost model also the new demands of their business.

Madelbrot by Frax

Madelbrot by Frax

 

 

 

 

 

 

 

 

 

 

 

 

 

Enhanced by Zemanta

The MS-Dos Phase Of BigData


The first developers of IBM PC computers negle...

The first developers of IBM PC computers neglected audio capabilities (first IBM model, 1981). (Photo credit: Wikipedia)

Most recently I was invited to talk to various CIO´s in the German context. The focus of my presentation was about the next years and what to expect from society, how IT changes everything, how to prepare and what to educate the employees on. IDC talked about the concept of the 3rd Platform. Thinking about the BigData hype while preparing the day, I was reminded on my times with MS-Dos. There was not only one approach to the “PC-Aera”. There was DR-Dos, Apple, still Mainframes etc.

One of the reasons MS-Dos gets more attention was the affinity to IBM and the combination of a device with a market place at this time. Also IBM let others open the space to develop software, add-ons for the PC etc. This generated momentum.

Seeing back the development to Windows 7, has taken a close combination of physics, like Intel, Software, like Microsoft and a change in the market.

The same we currently experimenting with cloud, virtualization and the future stage of an Software defined Datacenter, I guess we have long passed the MS-Dos phase.

So what is the MS-Dos phase?

Plenty of players, not a defined market, overselling of functionality, misleading abbreviations and very poor educated experts, best are from the vendors itself.

With all my talks to the CEO´s, CTO´s of industries, it has the same questions open:

  • What can I do with this technology?
  • Who can help me to build out the business context? and
  • Is this already time to invest into?

There is a clear answer, it depends !

It depends on the maturity of the IT department:

  • are they still maintaining the IT or do they drive innovative, business related IT processes.
  • Where is the infrastructure? Is the IT department still deploying Storage, Network and Servers, or they running a converged infrastructure with a services layer?
  • Do IT departments still focus on education on infrastructure or they hiring Data Scientists ?
MD Tag: MS DOS

MD Tag: MS DOS (Photo credit: shawnblog)

If you can answer this for your IT department, you are much closer to leave the MS-Dos phase behind you.

In the past, many business models have been established and run for years. This will be no longer true in many traditional businesses. Also this business is often in MS-Dos phase, but it will adapt thru the new market drivers, consumeration, urbanization, mobility, Maschine-to-Maschine decision processes and new adaptive computation processes which adjust the business to the demands of the users faster than realtime.

New concepts are unpredictable, like crowd sourcing, open source, raspberry pi physics, and machine learning. There is only one big truth, from MS-Dos to OS X it will take not 20 years, to will happen tomorrow.

Enhanced by Zemanta

My top 10 anticipations for 2013


The cloud will emerge in all parts of the IT

More services have been emerged in 2012, the adoption will drive more cloud, and more cloud offerings will parer. This circle will speed-up and drive more Enterprises in evaluating the next private cloud.

bigdata_network

The engine will be build in the manufacturing not at customer premise

In many enterprises 2012 has started a evaluation of converged infrastructure. Since most of the used components are standardized the „build on customer site“ or „do-it-by-myself“ will be more and more questioned by the CFO. Like in the past where whole servers where build by locals or enterprises itself, it is obvious that the purchasing department will have a different look when it comes to TCO. This will also start to include backup and security.  The order will be moving to workloads and software demands rather than core´s and PB of storage or Network interconnects. The VCE model will become the standard delivery methods for modern DC architectures.

Consummation of IT will drive more business adoption in the cloud

The DC will have more direct contact with the end customers. This was in 2012 one of the main drivers, build thru Google/Android and Apple/iCloud. This will move faster. The new paradigm enables enterprises to optimize cost and optimize the business model when direct talk to the end-customers. In 2013 we will see that this shifts also into computer-to-computer relationships. The interaction of b2b mash will lead to faster purchasing and information flow to optimize the business, and early adopter will be the winners. Key here is an agile IT infrastructure.

Cloud will feed BigData, BigData will enable more cloud

Each one is depended from another. With more cloud, the information streams are better to combine for BigData Analysis, with more BigData analysis and applications, more agile computing environments (called cloud) are necessary. This will be a big trend, but depending on the industry. New businesses will learn to use the information streams, new forms of analytics will be generated to enhance the processes and decision making progress.

Flash will be leading the storage industry

The last years flash was used as replacement for spinning disk. This has a huge impact in performance and utilization of computing power. Man-years of brain power where spent to develop smart algorithms to leverage the new TIER in the storage. This will continue to evolve and with the drop in price and increase of capacity the market will grow dramatically. However flash has not only be squeezed into  spinning disk, it also allowed to transform and implement new algorithm´s to utilize the power of flash. When IO´s no longer the limiting factor, CPU power can be used to leverage the performance in the storage system to do much more tasks which delivering high performance to the server´s.

The CIO will refocus on the core: the Information

With all the changes in IT, the transformation of the roles in the IT management will not stop. Since the DC gravitate to the information it is obvious to see that the CIO will be the master of the information. Not only the information processing in the own DC also the information needed form outside, like BigData applications, also informations given to outside, i.e. to customers. New technologies like the app horizon manager from VMware will support this transition and the CIO and team will be transformed into the information broker, security agent and consultant for the business lines.

Standardization will enable more productivity

One aspect of going from private cloud to hybrid is the standardization. Many companies, like Microsoft, Google or Amazon u.o. define the API and push them as de-facto standard. Experience showed that the early adopter often drive this, but run against a wall in a couple of years. Ethernet/IP and FC would not have been so  broadly accepted if there was not a standards body formed. We see currently various associations to  take on the role in the cloud, like the SNIA organization. This is the only way to help the DC out of the „do it by my self“ and focus on more business relevant tasks. The engine (converged infrastructure) will be developed and assembled in the vendors premise and the Enterprise DC managers can focus on the utilization.

The storm on clouds will drive orchestration

When, in the past, virtualization was introduced to customers, the VM landscape was able to be controlled by man kind. It is similar to the SAN in the early 2000´s when the storage array was still close to a server. This will continue to chance, when we will see 10.000 of VM´s in modern DC architectures. The orchestration will lead to the agility which is necessary to drive more business flexibility. Vendors, led by VMware,  will provide here more sophisticated solutions to automate.

Keys.

Keys. (Photo credit: Bohman)

Security and Trust will be in the middle of everything

Since Information is the key, like it was in the past but now it is a more open world, to secure is one of the key elements. The business will ask for answers, and companies like RSA will lead the way. Not only to secure also to trust other organizations is essential. With new regulations and demands of the business information has to be more trustworthy.

InMemory systems will draw investments from infrastructure

Since InMemory systems show more demand on the „main“-Memory than Hi-IO rates, it will re-architect the infrastructure in this area. 2013 will show if this new technologies will add new fields of applications or replace other. Technologies like GemFire, Hana u.o. will drive faster decision making and new infrastructure architectures. Combined with flash companies like SAP and EMC will drive the Industry here.

Enhanced by Zemanta

Keynote at the VMword 2012, with Pat Gelsinger and Steve Herrod


VMworld Barcelona

VMworld Barcelona 2012

Pat Gelsinger is on stage in is new role as CEO of VMware

Pat starts with an back view of the history. Waves of change has been the only constant at the IT industry. Led by IT innovations and driven thru technology. He talks about phases of IT maturity, reactive vs. proactive.
The transformation is in each layer starting with consumer, IT departments, people, operations and procedures. It began in various layers simultaneously and he talked about the infrastructure layer first of course.

From server to cloud.

This means from 25% in 2008 server are virtualized, today it is around 60%. The expectations that it will be in the near future more than 90% of all workloads virtualized.  This has also impact on the provision of servers from week, to days and in a view years to minutes and seconds or less. This only can be done by introducing a new paradigm automation

Next topic: software defined datacenter
Introducing the VMware perspective. Means that it is everything virtualized. Pat starts to talk about a huge legacy in the DC of today; the only solution here to end the dilemma is to abstract, pool and than finally full automate. Leading to the concept in manufacturing on „Just in time“ or all as a service.
The Software defined Datacenter is based on the vCloud suite, the basic element of the beginning of the journey.

Pat Gelsinger on vCloud Suite

This suite it is comprehensive, deliverers the highest performance, proven reliability, check it out.
Pat announced now that vRAM will no longer exist. based on the customer feedback VMware has decided that this is no longer the way to go. The new model is

  • priced per CPU
  • one easy solution
  • has no limitations

Back to the Software defined datacenter, diving into the management philosophy, more automation, less management, which will lead to IT broker of services also mean that the management must change.
The  policy based automation is the key to manage the future datacenter, eliminating time and human errors in the equation. This equals
service provisioning including vCloud Automation, vFabric Application Director,
operation management including vCenter Operation management suite,
Business management including IT business management suite.

The CEO now talks about Cloud Infrastructure around software, technology and architecture. This leads to the next area: How do I operate in the new World? its about People, culture and org, as well as Process and control and IT Business management
A new forum is build to shape the ecosystem, call „cloud ops forum
Moving on to multi cloud environment, like vCloud, physical, non/VMW and public.
Pat mentioned  PaaS is key represented thru cloudfoundry and automatic service provisioning DynamicOps, and software defined networking and security, keyword is nicira.

Addressing the multi cloud world
The VMware executive claims that it is ready for the open world.
Next high light is on applications, to move apps into new dimensions from vFabric on application transformation.
This all has been only be able to drive this transformation it was only be able thru the huge partner system.
Pat now moves the mic over to Chief Technology Officer Steve Herrod.

Future Storage Directions

Steve start to talk about the next generation  vCloud suite, and start to dig into. VMware, vSphere, virtualization, focusing on Exchange, share point server and SAP as well as Oracle.

What is a Monster VM, drive this to the edge.

  • 2011: 32 vCPUs, 1 TB per VM and IOPS 1 M per host, 
  • 2012: up to 64, 1 TB and more than 1 M IOPS per host.

So what about tomorrow’s applications, So what are they, in Telco, VoIP issues is latency, VMware is looking into that , based on vFabric, they look into the shared virtual memory, also HPC is coming to to virtualization, PaaS like cloudfoundry, introducing more on the Hadoop integration.

Enhanced by Zemanta

Learn from the past, think about the future, but live in the now !


When I prepared for my last tests at the University, one of my professors mentioned that they intend not to guide us to find good jobs immediately after the University, the key is that we as graduates have longer term chances to find and keep excellent jobs. At this time I thought it was stupid, my profession starts with my first job and not in 10 years.

Die of an Intel 80486DX2 microprocessor (actua...

Die of an Intel 80486DX2 microprocessor (actual size: 12×6.75 mm) in its packaging. (Photo credit: Wikipedia)

Time showed to me, that it is key to have both in mind when you start and it is essential to stay laser sharp focused.

Same in datacenters, when CIO´s get confronted with the new reality of BYOD, data deluge, cyber attacks and complexity explosion. Positions and processes years worked no longer seem accurate, personal gets confused, and the business units claim to have often more IT knowledge, since they use flexible services on the Internet. The CIO´s and IT managers I talked to often seek for external help, call consultants, which tell them that all has to change, processes has to be reworked and the HC has to decrease.

What a challenge!

Most people underestimate the finest attitude mankind has is: the best ability to change and adapt. Of course this is not easy, people like to be in the comfort zone, don`t like to get moved or learn new skills, move on and modify, even they do this all their lifes.

Most successful IT organizations I know have arranged with that to a superlative, taken the employees on the road and established a culture of change. As we learn from the past the change never stops, we moved from monolithic mainframes to mini´s to server to PC and now to tablets; from proprietary OS to open source from complex instructions CPU´s to RISC and back, from multiple CPU architectures, to multiple server to multiple core architectures. From ASCII terminals to mouse driven interactive PC´s to gesture tablets. From 640 KB RAM to Tb of RAM, from 5 MB Disk drive to 4 TB and now flash technology. So we know the past, we understand that telecommunication 10 years ago has been an enterprise asset and now is an outsourced, VOIP driven business.

So what will the future bring to plan, new devices, like the Google glasses, faster storage, from keyboards to voice driven interaction with computers, to InMemory DB build in CloudOS, from Applications to functions, from physical infrastructures topologies and silos, to software defined datacenters. From Application centric approach to information centric concepts, from blade farms to virtual workloads which independent from the blades will be executed?

We will see computer talk to computers, make decisions for us, devices will generate information, which only relevant for seconds. Batch processing will become too slow to keep pace;

This image was selected as a picture of the we...

This image was selected as a picture of the week on the Farsi Wikipedia for the 13th week, 2011. (Photo credit: Wikipedia)

it will be replaced thru bigdata algorithms.

On the consumption model we will purchase only what we consume, vendors have to deal with end customers enterprises will move legacy workloads to specialized and focused workbenches, all information has to be secured and trusted to be transferred.

So the key for the future will all about information, generates, capture, store, manipulate, destroy and analyze, we call it the gravity is on information.

Based on that, CIO´s and IT managers have to and do act in the now, to prepare. Build an agile infrastructure, make our investments in the Datacenter on skills, new technology, information and big data, that supports the workload management and secure the information, which this is critical to the enterprise or the privacy of an employee or customer. Investments in IT of an enterprise should help to build and prepare an agile datacenter around the information to be ready for the near future and bejond.

Enhanced by Zemanta

The DataCenter Evolution


In the last decade the DC has taken major reworks and design changes. The last years the evolution was marked thru the revolution on the market. Disruptive technology shifts and new classes of devices have changed the demands of a classic DC. Still most of the money is spent to keep the lights on. On average only 25% of the budgets of a datacenter is kept to new business and improve the situation.

The DataCenter

(c) by presentermedia

So what will happen ?

First of all there is not one answer to this question, but we can try to look into the future challenges of IT businesses:

The world will continue to generate huge amounts of informations. Consumeration of IT will continue at a faster pace, with new devices appear. Regulations will change to adapt the new reality of IT, like privacy by default etc. Realtime of Information will overcome, data collection for later processing Security and Trust alliances will be more necessary since Information is more volatile.

In 1999, when I experienced the first SAN implementations many CIOs told me that there will be a limit on data storage since the complexity become too much to handle in the DC with two or more networks. Looking back the technology vendors and IT departments have established a quite good understanding how to train and build out teams of architects for the SAN Networking in parallel with the IP Networks. Now the complexity in DataStorage raises and DC managers think about the way to deal with that. Future will tell, but my view is that the networking will come together with the SDN concepts where the physical network is decoupled from the logical view. In this case the SAN and IP can be managed from one point. Further more with the appearance of the Virtualization the future of the ressource management will be come together. When observing the technology development and the industry demand on information management, the cloud burst and the consumeration of information management the way leads to a generic ressource in IT which is assigned a task. In otherwords, why having a storage device, a computing device or an networking controller or even a PC. it is defined by the demand. There will be an efficient device for each task in the first step, but I strongly believe that in a couple of years the physic will no longer be the architectural definition of the DC. It will be the software.

The Earth

(c) by NASA

So why would a CEO today think about putting money into an own DataCenter or IT-Department ?

Most probably to gain advantage of the competition in having applications bestir customized to the business process. Today I found very interesting that CIO´s get asked by the business to challenges with the rest of the world. It is hard for them since the internal BU´s often not compare fair. The “good enough” principle will only work if this is also internal applicable.

Security demands, internal processes and other guidelines drive the IT-department often in additional cost.

So the question should not be “do I need an own DC of IT-department”, it should be the question about responsibility on the information. There is a reason why the DC is called Datacenter! and the CIO os call “Chief Information Officer”, not “Chief Datacenter Officer”. In this context, it is obvious that the DC no longer is necessary in physical form to exist, it is more the collection of services and Information flow the responsible manager should look for.

Finally the datacenter will become the Touring Maschine !!

Enhanced by Zemanta