Topics:
Q&A: Tom Simmons on telework
The mother of all gridlocks is set to descend on Washington, D.C. this Monday and Tuesday with world leaders converging here for the 2010 Nuclear Security Summit, at the Washington Convention Center.
The federal government will be open throughout the entire ordeal of street closings and public transit rerouting. The Office of Personnel Management says it "strongly encourages" personnel to telework.
FierceGovernmentIT spoke with Tom Simmons, a vice president for Citrix public sector, shortly before the summit about the state of federal teleworking. What follows is a lightly edited transcript.
FGIT: There's been a lot of talk about teleworking, over the last few years. How prevalent is it truly?
Simmons: It's not extremely prevalent across government. I think we're a lot further today than we were three years ago when Rep. Frank Wolf (R-Va.) put it on the front page and said he was going to tie appropriations dollars for those agencies he sat in oversight on to implementing telework plans. I think in those agencies with the funding, there's been planning to the point of implementation. But implementation across the board is still probably well less than 5 percent of the targeted user population.
I know the most recent administration targets were 50 percent [c.f. OPM strategic plan, 2010-2015 (.pdf)], so there's a long way to go.
FGIT: Why is it that telework really only encompasses 5 percent of the targeted population?
Simmons: I think the largest issues for government managers is the ability to put management parameters around outputs, and results, around measurable things rather than face time or time in attendance, or stuff like that. I still believe there's a general--mistrust is too strong a word--but, there's a general feeling within government that folks who would be at home would spend more time doing chores around the house than actual work. Unless and until they can get to the point where that work becomes measurable, the challenge around telework is going to persist.
In private industry, especially in a sales organization, it's easy for me to set management parameters. I can manage the number of phone calls made, I can manage the numbers of proposals submitted. Ultimately, I manage revenue contribution by everybody on my team. So, telework is fairly easy for us to implement. Compare that to the typical government agency whose mission is to provide services to the citizen or services to other government agencies.
I think the burden is on middle-level management to come up with the measures of outputs or results that would make telework not just an acceptable, but a preferred way of getting things done.
FGIT: What kind of metrics could be realistically implemented?
Simmons: When you look at the agencies that have been very successful at implementing telework--
FGIT: Such as?
Simmons: The Defense Information Systems Agency is one. They probably have the largest percentage of populace teleworking on a regular basis. Charlie Croom [the former Air Force Lt. Gen. who lead DISA from July 2005 through July 2008] said, ‘We're going to move, telework is going to be an essential part of our re-location strategy because I don't want to lose all the intellectual assets that I have in Northern Virginia all of a sudden when we open up offices outside of Baltimore.'
He looked at job functions, and said ‘Alright, with each job function, let's identify three or four metrics that we can measure, that we can send people home two, three days a week and still measure how productive they are.'
He gave a briefing and what used to be called IPIC in Orlando two years ago [now the GITEC Summit] and said he even looked at his administrative assistant's role in the organization, and said ‘Does she really have to do this from a desk right outside by office? Technology allows her to answer my phones wherever she is. She can manage my calendar, she can update my daily system routine from a computer from her home.' He even went so far as to personalize the implementation of telework to his secretary and came up with the kind of things she needed to do on the days when she wasn't in the office, and days when she was, that might change a little bit, because you might get some of the things done in person, like re-stocking a supply shelf, things like that.
FGIT: Well, what kind of metrics can be implemented that don't seem like they come from a Maoist collective trying to measure every single second of your life?
Simmons: Yeah. Generally, depending on work function, systems administration, customer service--you can measure phone calls made, you can measure results of those phone calls, whether it's memos to file or tasks delegated or tasks completed. You know, I've been a teleworker since 25 years ago, and I find that both as a home worker and as somebody that manages people there, we get far more out of our people who are able to work from home. Because that work environment is always there, always available.
That's the concept we need to get to with government--is to say, don't look at the negative side of it, ‘How do we keep the home worker from doing laundry or watering the grass during the day,' to, ‘Instead of just commuting in and just working from 8:30 to 5, I might get from 7 until 11, doing emails, catching up on administrative chore work and stuff like that.' It's that entitlement enablement approach that I think has proved so successful at DISA--and at places like GSA, where they've also done a pretty good job of implementing telework, because there's that trust for the employee that says ‘Get the job done. If you can get the job in four or five hours, that's great. If it take you seven or eight or nine, well, working from home you can work that into a 24 hour day.'
FGIT: Must there be any policy changes?
Simmons: I think the biggest policy concern that is preventing change is if you look at telework policy, it talks about how you will manage your day, and how you will conduct business during business hours and not do some of those personal chores around the house. You got policy that assumes the worst, and that's kind of typical of government policy to begin with. Instead of looking at ‘How do I enable,' we often fall back into ‘Alright, these are all the boxes I have to check, the parameters I have to put around this so that I am not vulnerable to employ action or lawsuit or union kinds of things.'
FGIT: What about technology? A home-located computer not synched up with a work computer?
Simmons: With technology today, that really is, or can be made, a non issue. Most organizations have a set of specs that you will adhere to in order to gain access to an organization's network and information technology resources. In the bring-your-own-computer concept, it's not so much about the receiver, the device that you use to access, it's more how you configure the information technology resources for delivery to any computer.
And with virtualization technologies, especially around desktop virtualization, all you really need is a browser to access a virtual work environment that's hosted within the organization, and you can establish the kind of policy and procedure that says, ‘The data will never leave the data center.' If I'm a financial analyst, and I'm looking at agency data, all I'm doing is looking at a picture of that data cross an Internet link. The data itself, the PowerPoint or Excel application used to manipulate that data is running on a server. It's my iteration of Excel--I've got it personalized the same way I would if I were using my desktop at work--but there's no data going across the wire. There's no number crunching or data crunching taking place at the client side of the equation. All that's being done in the data center. And with the advent of cloud computing, that becomes an even more widespread, ubiquitous opportunity to provide remote access.
FGIT: Does that assume everybody at work has a thin client?
Simmons: If that were the way you implemented a desktop infrastructure within an organization, you get the benefit of doing the exact same thing as at home as at work. It minimizes the transition.
FGIT: So you do need a uniformly thin-client environment?
Simmons: Thin client is just a device. You can run a virtual desktop on a dual core Pentium processor, big beefy Dell or HP desktop machine. It's just a matter of where is the processing and where does the application and data reside. The beauty of virtualization is I centrally manage all that stuff in the data center. Whether it's a desktop PC, a thin client, a notebook or netbook--or, in today's world, an iPad or an iPhone--via a browser interface, regardless of the access device. I can address all the security issues that the government has historically done by locking down a government-furnished piece of equipment, by locking it down in the data center, but I provide access to any of the available devices that my employee might had and I put restrictions on how that data can get used. I don't allow local print, I don't allow local copy--I can't copy data cross that link to a thumb drive and move it around, if security policy says ‘You shall not do that.'
FGIT: Thin client isn't the most popular thing in the world...
Simmons: That truly is a cultural thing. I worked at Compaq for eight years, and we probably did most of our desktop and notebook revenue not based on the technical requirements of what the user was doing, but more on the status symbol and prestige that a new PC every three years, upgraded with the most memory and the fastest processor and the full suite of plug-in accessories [renders onto the buyer]. The buying motivation for desktop PCs or notebook PCs is less, ‘That's what it takes to get the job done,' because in a world of thin client computing or any client computing, we can show--and we have--that the real meat in an IT implementation for application support is in the servers.
FGIT: But you're subject to the vagaries of the network--a single failure node for many employees.
Simmons: In certain kinds of network configurations or infrastructure configurations, you're absolutely right. But in most cases, you don't look at Google or Amazon or MSN as a single point of failure kind of an approach to accessing applications.
FGIT: Gmail still goes down. Is it an annual thing, Gmail goes down?
Simmons: Probably. But compare that to the uptime of an agency's implementation of Outlook. Is Gmail more reliable, or less reliable, then an Outlook implementation?
FGIT: Google also owns half the servers in the world.
Simmons: Hm, mmm.
FGIT: And even they fail.
Simmons: Generally, what we're seeing in government datacenter policy today is that they're both expanding their own government-owned data centers, and out-sourcing either augmentation capacity or COOP capacity from third party providers. What we're seeing in most mission critical applications and mission sensitive applications, to include email, is a continuity of operations and disaster recovery plan. And when we build those on a physical and virtual platform environment, I can build an Outlook workload, I can install that workload on a set of physical servers in my data center, or I can virtualize those servers so I can take care of peaks and spikes in capacity demand. And once I've virtualized that workload, I can move that workload around within my infrastructure. So if I have a data center that's in Washington, D.C., that's running Outlook right now, and I've got, let's say 60 percent of that workload running on virtual machines in the datacenter and that datacenter fails, I can move or re-provision servers in my facility out in Herndon or out in Culpeper.
I might see a minute or two blip in my ability to refresh or reload an email that was on the screen, but my backend structure is all synchronized. It's the early stages of a private cloud, to do that kind of compute thing.
We don't really promote or advocate thin client. If I need a device that has local compute and local storage capacity because my network is subject to outages and the like, then I can continue to use desktop PCs or notebook PCs. But that doesn't preclude the use of a virtual work environment. I can provision across the network into a local virtual machine the application, the data around that application and utilize the local computer process of a laptop.
A local machine on the laptop can connect with the work environment of an organization--I can use the local compute capacity of the laptop. I can, if security policy dictates, use the local storage and print, all that kind of stuff. But the environment that I'm tying into is a hosted environment in the datacenter, and it's all secured by the firewall and managed centrally.
FGIT: The datacenter is taking periodic snapshots of the local device?
Simmons: The device is taking snapshots of what's happening in the datacenter, so if my connection goes down, I still have the application code and the data associated with the applications so I can continue to work locally until the connection is restored.
Usually it's not the technology that's going to be the impediment to a successful telework implementation.
Generally, when you're talking about a teleworker you're talking about Microsoft Office of similar kinds of office productivity operations. You're talking about access to legacy mainframe or hosted applications that run in the data center. The client there is really nothing more than a communication terminal. You may have some homegrown applications that do in some situations require a client to be loaded traditionally on a client device. But when you look at what those are--HR, for example. HR in government is either SAP, PeopleSoft or Microsoft apps. Generally, any time you're doing a remote access to that, you're doing it via the Internet implementation or the web implementation of that. MySAP or the like.
So, the dependency of the application suite on being installed on a client device is less today than it was even two years ago. And it's migrating more to a centralized hosted environment.
Related Articles:
Federal telework hurt by IT security, budget pressures
OPM chief endorses telework




Comments