Wednesday, April 29, 2009
Pointers for IT Business Continuity & Pandemics
While the chatter in the market is all about connectivity, I'll point out that a laptop and a VPN aren't sufficient -- there are managerial and organizational issues to take notice of.
Now, some jobs don't need you to "be in the office" and in fact, sometimes being in the office is plain wrong. For example, if you're a salesperson, you're probably in an office 10% of your time, if that much -- so why do companies pay for that permanent office space? The other example of when going to the office is wrong is during a potential Pandemic... As we've seen this week in Mexico, the government has already taken steps to reduce congregating individuals by closing-down the schools. This could easily escalate into voluntary (or mandatory) steps whereby adults need to stay home from the office.
But how would this affect business operations? Most business continuity / continuity of operations (BC/COOP) focus on keeping the IT portion of the business running... but what if there aren't any people to run them?
Lots of content is now being written by industry analysts; In a recent Forrester Research Blog, Stephanie Balaouras writes "Swine Flu? What It Means For IT Professionals"; they cite a recent joint Forrester and Disaster Recovery Journal survey regarding BC strategies.
Gartner Research has also put out a press release "Swine Flu Is a Reason to Act, Not Panic," and Networkworld has a pretty insightful "10 Tips for swine flu planning."
But let me share one aspect that these high-level recommendations don't touch-on as they should: Managing the organization when it's distributed and/or remote.
Getting all of the VPN and laptop technology into the hands of workers is one thing; keeping them functioning as a cohesive group with management oversight and direction is another. If people end up working remotely for 2+ weeks, this becomes a critical issue.
I've spent some serious professional time examining this, particularly with one of the leaders in the field, Sun Microsystems... where a significant fraction of employees are remote/virtual workers. Sun has a mature approach to implementing this, called Sun OpenWork. Check out a very excellent whitepaper on "Managing in a virtual organization"
You should also check out TeleTrips, who specialize in consulting for distributed/virtual organizations, telecommuting programs, scenario planning, and online tools. Whether your enterprise wants to invoke a mobile work program, or just have a BC/COOP plan in place should a disaster occur, these folks help examine work and management practices, facility locations and technology infrastructure, and then put in place the appropriate programs & training.
It's all very different from a laptop and a modem. And should a natural disaster strike (esp. here in sunny CA) people might not be able to, or might be told not to, come to the office. Be prepared.
Monday, April 27, 2009
Who's using Infrastructure Orchestration in Healthcare?
It's the approach that HP (w/Matrix Orchestration Environment plugins), IBM (Open Fabric Environment) and Cisco (Unified Computing System or UCS) have entered into recently, and where firms like Egenera (with PAN Manager software and the Dell PAN System) have been selling for some time. The beauty of infrastructure orchestration is that it abstracts-away all of the "plumbing" of the Server such as I/O, networking and storage connectivity -- which makes it an absolutely *ideal* complement to virtualization. So this technical approach is gaining greater installed-base every day.
Taking the conversation from "cool stuff here" to "it's in real use in the real world", I thought I'd give some examples of users of Egenera's PAN Manager software in the Healthcare industry:
Cambridge Health Alliance:
Cambridge Health Alliance chose PAN Manager to manage the infrastructure for their ambulatory product suite, in a strategic initiative to automate its ambulatory-care environment. Over five years, the Alliance expects to save $2 million, including $1 million in initial capital costs. Equally significant, the software is reducing system administration requirements—enabling Alliance IT professionals to focus on activities that add real value to the user community. I like this part: “If we’d purchased any other platform to support the... applications, we would have had to hire more system administrators. The simplicity and automation of the Egenera system take the place of two people.” Chief Information Officer, Cambridge Health AllianceEmory Heathcare:
Emory Healthcare moved mission-critical applications from proprietary UNIX® and mainframe platforms to a virtualized infrastructure based on industry-standard servers and PAN Manager software. Benefits included improvements in system administration, TCO, utilization, and provisioning time. A system administrator can configure and allocate a virtual -- or physical -- server in minutes. In addition, multiple operating systems and OS images can be run on a single server, enabling IT to create test systems on servers otherwise used for failover. PAN Manager security facilities also help ensure that patient records are preserved in accordance with government mandates.Metavante Healthcare:
With annual growth of 35 to 40 percent, Metavante Healthcare Payment Solutions needed to upgrade data center infrastructure just to keep pace. The company chose PAN Manager as an Infrastructure Orchestration approach for its unique virtualization capabilities. PAN Manager had quantifiable improvements in availability, performance, flexibility, management, and cost savings. Again, I live for this: “I was looking for a system designed from the ground up to overcome the limitations of standard servers… We didn’t see an architecture like Egenera’s from anyone else.” Vice President and CTO, Metavante Healthcare Payment SolutionsSCBIT (Schanghai Center for Bioinformation Technology):
SCBIT chose PAN Manager infrastructure orchestration for its ability to simplify consolidation, virtualization and management; to reduce application time-to-market for applications; and to lower data center costs. PAN Manager provided the flexible allocation and repurposing that SCBIT required: The agency can run any of its 10+ applications on any server at any time. PAN Manager also enables SCBIT to make every application highly available at virtually no cost and provides a unique N+1 approach to disaster recovery. Their evalutation also showed performance advantages of running Oracle 10g on servers with PAN Manager, versus traditional systems. Cool.
Overall, it's nice to see that this technology has major footing already. My belief is that as Virtualization becomes more pervasive, and as data center technologies become more complex, we'll see this infrastructure orchestration approach begin to displace "traditional" approaches for IT Management. It's simple, elegant, runs on standard x86 boxes, and provides broader reliability than typical clustering solutions.
Keep your ears open. for more.
Monday, April 20, 2009
Who's using Infrastructure Orchestration in Finance?
Indeed, there are a number of similar technologies out there (Cisco with its UCS, and HP with its own offerings). But sharing actual users/uses, I believe, takes the conversation from "hey, cool stuff here" to "gee, it really is useful in the real world."
The beauty of infrastructure orchestration is that it abstracts-away all of the "plumbing" of the Server such as I/O, networking and storage connectivity -- which makes it an absolutely *ideal* complement to virtualization.
The results of this approach is that physical servers can be repurposed easily, regardless of whether they're running physical or virtual software. And fast repurposing means you can deliver instant High Availability (HA), entire environment disaster recovery (DR), and near-instant scaling (capacity-on-demand). In fact, even without VMs, some consolidation is possible by being able to use the same box for different uses at different times.
Within the past few years, a number of financial-services firms have adopted this approach using Egenera's PAN Manager software:
Commerzbank NA:
After the 9/11 terrorist attacks, Commerzbank NA launched an initiative to enhance its disaster-recovery strategy. The goal was to deploy a pool of virtualized servers that could run UNIX or Windows on demand, to complement the flexibility already achieved with virtualization on the storage and network sides.Standard Chartered PLC:
Today their DR site not only ensures business continuity, it plays an active role in daily computing requirements—notably improving utilization. Plus, Commerzbank NA has consolidated 140 legacy servers into 48, slashed server-configuration time from two days to one hour, and reduced floor-space requirements by 60%. Niice..
IT architects at Standard Chartered PLC decided to centralize data center ops for core retail-banking application, available to customers at 1,200 locations worldwide. They selected Egenera's selected PAN Manager to do the job. As a result, Standard Chartered has cut total cost of ownership in half compared to their previous proprietary solution and can now bring a new country online in nine days rather than 45 days... as estimated for its legacy architecture.Farm Bureau - Western Computer Services:
Western Computer Services, Inc. (WCS), a wholly owned subsidiary of Farm Bureau Financial Services, deployed Egenera's PAN Manager software the foundation for its new service-based architecture, serving multiple Farm Bureau Financial Services companies throughout the United States. The Farm Bureau chose Egenera to power the delivery and management of Web-based insurance services to thousands of personal and commercial insurance customers.Investment Banks....
Egenera has a bunch of Investment Bank customers - in uses for HA, DR and repurposing, for applications from order management, order routing, and other client services. But these guys won't let me use their names :(
Friday, April 17, 2009
Postcards from SDForum's Developer's Conference
The day opened with James Staten of Forrester Research - giving his usual riveting, insightful view on clouds, cloud adoption and directions. I think he also was able to add some sobriety to the hype, pointing out that Infrastructure-as-a-Service was the technology most likely approach to mature first. He expects to see "cloud hype" to die-out around 2010.
James was followed by a panel hosted by Chris Yeh, focusing on new software business models. Pretty lively discussion about how packaged software is moving to subscription, how SaaS is re-making how software is consumed, and how simple financial models will help shape how products/services are packaged and priced. Frankly, the internet is changing all types of business models... e.g. a question from the audience focused on traditional advertising business models, where "classifieds" were advertising in a newspaper where "news" was the content. But with the advent of Craigs List, "classifieds" *is* the content. Hmm.
Now a shift: a panel on Mobile development - Panelists from Nokia, Sun's Java division, iPhone developers, etc. One really interesting insight: think of the phone application market as "verticals" and "horizontals". While there are a few million phones that are "vertical", i.e. Blackberry, iPhone, etc., there are a few *billion* phones globally that are more basic, but where there is a more massive market to write to. The horizontal market for mobile dwarfs that market that we think of as the "advanced" iPhone market!
After lunch, we got a really cool and riveting presentation from Clara Shih from Salesforce.com on "Understanding the Facebook Era" Her premise: Facebook is CRM for individuals. Ergo, there is a natural connection between Facebook and SalesForce.com... (Faceconnector) Also, there is the need for "online identity" and Facebook is making a play for owning that 'credentials' space (Facebook Connect). But with online identity, there is the potential for massive online data mining and demographic research.
My own panel was moderated by Chris Preimsberger of eWeek - focusing on cloud infrastructure. Hamid Pirahesh of IBM led-off with a really elegant 4-quadrant perspective of Traditional vs. Cloud models, and internal vs. external models. Lots of good conversation about why "cloud" infrastructure differs from traditional infrastructure.
Finally - the VC panel... Accel, Draper Fisher Jurvetson, Hummer Winblad, and Benchmark. First question off-the-block: "Where are you seeing the growth?" Answers: (a) small-budget items that can be purchased in bite-sized pieces, (b) IT infrastructure equipment that saves money, (c) data analytics, cloud and cleantech. Mostly pretty bullish on the startup market, apparently lots of series-A and B happening now. And lots of really good advice for wannabe startups on how the VC process works.
Monday, April 13, 2009
Private Clouds are Real and In Use
Turns out this is really happening, and I'm thrilled. In "Why 'Private Cloud' Computing Is Real -- And Worth Considering," InformationWeek reported that this isn't your father's Oldsmobile... it's real and it's different:
...The Department of Veterans Affairs has deployed a small internal cloud. It wanted an early-warning system that could analyze data from its 100-plus clinics and hospitals and spot outbreaks of infectious diseases, and it had to do so on a tight budget. The project, dubbed the Health Associated Infection and Influenza Surveillance System, was built on six standard blade servers with converged network and storage I/O. The CPUs can be managed individually or as a virtualized whole, with workloads shifted and capacity summoned as necessary.
The system runs Egenera's cloud management software, PAN Manager, which manages I/O, networking, and storage for the servers as a logical set. It can execute several applications, while always having enough horsepower to do its main job. The system's Dell blades and storage can be virtualized as a pooled resource in such a way that processing power can be devoted quickly to the VA's cloud, its highest-priority task. In many ways, the VA's new system anticipated Cisco's recently introduced "unified computing" platform, a virtualized, multiblade server chassis with converged I/O that Cisco touts as just the thing for cloud computing.
I've spoken with the VA's CIO; they're running both physical and virtual applications (i.e. physical Oracle and virtualized services) but want to be able to scale transparently, with extremely high levels of availability, etc. Plus, like many IT Operations, they have trouble anticipating end-user demand -- so they require instant "elasticity" in the system. Hence, the "cloud" model for IT operations fits the bill. While they don't call it that ("cloud" seems to be the sexy term right now) it's how they're operating.
My suspicion/hope is that as this simplified model for internal IT matures, more IT operations folks will see the light (through the clouds).
More later.
Tuesday, April 7, 2009
Egenera creates physical & virtual servers transparently
PAN Manager is the "secret sauce" Egenera has been using to manage its own hardware for years, and recently OEM'd to Dell as part of their Dell PAN System. The software has always had the ability to provision physical and/or virtual servers with mission-critical levels of availability, including networking, I/O and storage configurations.
Now Egenera has bundled the virtual server provisioning option in at no charge into the core management product. The technology is based on the latest Citrix (Xen) technology.
That means that Egenera's own hardware -- as well as Dell Blades -- not only operates with a "unified computing" architecture (as some would call it!), but that operators have the option to create physical or virtual servers on-the-fly, as conditions require. And, regardless of the type of server (or software workload) the servers are protected with HA and DR, even across remote locations.
For Dell, this gives them a play in the mission-critical computing market, with a way to either embed VMs within their bladed environments, or to support other virtualized environments like VMware or Microsoft -- all with similar five-9's of availability.
Using the system is super simple and elegant. You'll never need to re-cable a server again, and never have to worry about I/O, network configuration or DR configuration. IMHO it will be the infrastructure model the industry will migrate to.
Monday, April 6, 2009
What IT Can Learn From the Bean Counters
The ITFMA meeting is just what you'd expect: lots of Sr-level folks, mostly finance, assigned to supporting IT organizations, data centers, etc. Some were technical, some not. While there was a good turn-out (and there is each year) I was surprised that it wasn't larger. This seems to be *the* place where the business-side of IT is examined -- and is one of the best-kept secrets in the industry. I certainly hope the conference continues to grow.
The topics - and questions - were telling of the state of IT... from a business perspective. Follows are a few of my takeaways from some of the talks/presentations:
- Data centers & energy efficiency: It was surprising to me how many companies admit that their data centers are 20+ years old -- tacitly acknowledging the inefficiencies they're coping with. Just as many were planning migrations or upgrades, even in this economy.
- Chargebacks: A surprising number of firms are using chargebacks, or planning to. Overall the industry average across IT is somewhere in the 20% range. But the people here seemed more likely to have already adopted it. And, with the advent of shared IT (i.e. virtualization) the trend toward chargebacks, and chargeback tools, seems to be growing.
- Software use & metering, + Cost allocation: As above, this continues to be a hot topic -- if you're doing chargebacks, you need to monitor how (and by whom) software is being used. So, lots of interesting topics on licensing, metering, etc., and lots of new vendors pushing solutions.
- SW asset management: Another subset of above. But now, with virtual servers everywhere, outsourcing parts of applications, etc., the topic of asset management seemed to be in a number of placces
- IT Process & Financial management: Dovetailing ITIL management processes with financial management. This isn't trivial for large organizations. Running IT involves capital improvements, problem management, change management, new service delivery, service retirement, etc. Finance has to work alongside all of these processes while the plane continues to fly.
- SaaS and Cloud: As you'd expect, there were a number of talks regarding just what Software-as-a Service is, as well as what cloud computing promises. Not so much from a technical perspective (remember the audience) but from a business, compliance and security perspective. Ultimately -- from a business perspective, the Cloud model represents a new business outsourcing model which is very convenient for certain classes of applications.
• IT Chargeback & Activity Based Cost Management Conference
• IT Expense & Asset Management Conference
• ITIL Financial Management Conference
• IT Telecommunication Financial Management Conference
Wednesday, April 1, 2009
eBay Is Really a Huge PaaS Provider
eBay has announced its Selling Manager Applications Beta, which is really a number of hosted services, each complete with its own API.
eBay Selling Manager Applications allows developers to help professional sellers manage their online businesses within their trusted My eBay experience. Professional sellers are like any small business, and quality selling management tools embedded into the Selling Manager experience will make them more efficient and profitable. Sellers will be able to find and easily subscribe to applications through an applications directory.Some of the APIs include
- Shopping
- Merchandising
- Feedback
- Client Alerts
- Platform Notifications
- Large Merchandise
- Research
eBay's roadmap page and current .PDF. eBay is also implementing these APIs using the gadgets specification as defined within OpenSocial.
What appeals to me here is that, unlike "raw" hosted compute services (a la AWS), and unlike raw platform tools (a la Google App Engine), eBay has chosen to host core services for use (and re-use via mashups) by others. eBay is putting its core IP on the web for others to use.
Not convinced at the potential? Consider the following thought experiment: What if eBay provides a highly-reliable, generalized trading application that matches Puts and Calls? What if a public company were to allow the engine to match and record trades of its shares using that engine? Could it eventually usurp the NASDAQ engine?
Now think even a bit more differently -- rather than a service that matched simple pricing (for shares or for simple commodities), what about a service that matched trades for more complex variables? For example, different grades and granular chemical compositions of crude oil, complete with amounts and delivery dates? Jobs and job seekers? Singles looking for dates?
eBay clearly has the opportunity to be more than an online e-commerce site. It could be the "Exchange Platform" in the sky. Very cool possibilities, well beyond the simple way we're thinking of cloud services today.