Good post by Alan Murphy. Worth a full read, particularly his take on "the good", "the possibly great, possibly not", and "the terrible." His call for an enterprise-class virtual switch certainly resonates with me.
So my recommendation to Cisco would be: Stick with what you do really well, L2-L4 IP networking, and let the people that do storage networking well do storage networking. By all means extend that L2-L4 knowledge and expertise into the virtual platform arena by working with VMware on building a usable and robust vSwitch, but stop there. We need a virtual data center platform that includes an enterprise-class virtual switch. But on storage…there’s already going to be a push towards storage VM appliances in the next few years; let them fail on their own without you mudding up the waters by trying to manage the storage network underneath that. ...
At VMworld 2008, VMware made a lot of their intent to foster the ecosystem of partners around the management of virtualized datacenter infrastructure aka Infrastructure vServices ... vNetwork, vStorage and vCompute.
In this interview, Carl Eschenbach (EVP, WW Field Operations) emphasizes the vStorage. Without taking the issue head-on, VMware seems to be sending a message that, under the newly arrived executive, the company will do a better job. It would appear that they are addressing the company's reputation for ill-treatment of the smaller members of this community.
VMware earlier this month introduced an initiative to help virtualize their server, storage, network, and application environments to take their computing to the cloud. However, questions remain about solution provider opportunities as well as whether part of the storage business will be supplanted by new cloud features. Carl Eschenbach, executive vice president of worldwide field operations, recently sat down with Joseph F. Kovar, senior editor of Everything Channel, to answer these questions. ...
In this post by David Davis, there are a number of good observations and a couple of issues worth pondering.
First might be what it means to "manage and monitor" virtualized infrastructure. If Packttrap or Solarwinds permits that part of the IT organization responsible for the network to manage virtual network componentry, at what point do they pull it all together into a unified view of "the network"? How does this happen without the network guys encroaching on the territory usually reserved for the "server tribe"?
One might argue that Cisco's Nexus 1000V recreates for the network organization a distributed virtual switch that, for all intents and purposes, acts like and is acted upon in a manner with which the network guys are familiar. The question will be whether this is ultimately a case of defining the use of new, disruptive technology (server virtualization) in terms of the old established technologies (physical switching a la IOS). (You can see one point of view here, in which Davis sets out his take on the 1000V.)
As for the challenges he lays out ... well, we think we know the answers to some of this, and intend to prove it. Answering the question about whether to support VMware ESX only, or other platforms is an interesting commercial decision for most players and bespeaks an understanding of the customer base. (When does Hyper-V have enough of a market share to justify the attention? Do customers have a requirement to manage both ESX and Hyper-V in the SAME virtualized datacenter?)
More and more of the typical “physical computer” management & monitoring tools are being retooled to manage the new virtual infrastructure. I have talked with both Packettrap and Solarwinds and both have rumored that they will soon offer versions of their well known network management tools that will now recognize, not only network devices and physical servers, but the virtual guest operating systems that are on those physical servers.
For example, your network management & monitoring tool could query either each individual ESX server using traditional SNMP calls or it could query the VMware Virtual Center server using VMware’s API to obtain an inventory of what virtual guest is on what physical server, performance statistics for both host and guest systems, and status of guest systems (ie: which are powered on or off).
There are a few challenges that these vendors face:
* do you go directly to each virtual host or to a centralized management server? * do you support only VMware ESX Server or do you try to support other virtualization platforms such as Microsoft’s Hyper-V? * how do you learn about guest VMs that have been “VMotion’ed” (for lack of a better term) from one host system to another? And what about the performance statistics when the storage for a guest is “SVMotion’ed” from one datastore to another?
So, “stay tuned”, as they say, for physical tools to now recognize the virtual world. And, if your vendor isn’t already doing this or doesn’t have plans to do it, I recommend that you pressure that vendor to make their product “virutalization ready” (or else you may have to go find another vendor).
Stephen O'Grady, one of the indominable Redmonk analysts, has a great response to a recent interview of Richard Stallman as reported by the Guardian.
Over the past few days, the noise generated by the interview, in which Stallman refers to cloud computing as "stupidity" and "marketing hype" has been reminiscent of religious arguments or ideological debates on free market economics in the US Congress.
However, I'd really have to recommend O'Grady's thoughtful treatment of the issue, and his (deservedly respectful) response to David Berlind about Free Software / Open Source Software. O'Grady correctly (IMHO) establishes the notion that F/OSS has a huge impact on SaaS, PaaS, and IaaS offerings. In part, it's because the vendors of proprietary softare, platforms and infrastructure have been happily minting money based on a classic computing and product economy (vs. a service economy). In part, it's because the F/OSS arguments really DO hold for technologies in which there exist reasonably well defined reference implementations ... operating systems, web servers, application platforms, data base management systems ... that lend themselves to standardized interfaces or by virtue of market dominance (like MSFT's Exchange) have become a de facto standard.
I'd recommend watching the likely exchange between (and around) O'Grady and Berlind. It's likely to be well worth the attention.
All of which begs the question of whether or not cloud computing is a clear and present danger to either or both of the free software or open source movements. From his comments, it seems reasonable to deduce that Stallman does so regard it. As does one of the smarter journalists you’ll meet: one David Berlind.
Personally, I disagree. To quote Austin Milbarge, for once I am in complete agreement with my partner: far from marginalizing F/OSS, the cloud may well prove to be its most significant benefactor.
Consider, for a moment, that the F/OSS’s impact on the PC has been, desktop marketshare-wise, neglible. And that while successful in the server market, it is hardly dominant. Yet in the cloud, F/OSS is at present the rule, rather than the exception.
Amazon? Built on Xen, hosting only (to date) open source operating systems (Linux and, in alpha, OpenSolaris). Google App Engine? Only the folks in Moutain View what, precisely, it’s composed of, but there’s little question that the foundation is open source. The language (Python) and the primary framework (Django), meanwhile, are as well. And so on. Some of the success of F/OSS within the cloud is doubtless attributable to the specific players. By Ray Ozzie’s own admission, Amazon - a shop with deep experience in open source technologies - simply took the cloud market far more seriously than did Microsoft. If the roles were to be reversed, the makeup of the cloud might look radically different.
But the foundational role that F/OSS plays within the cloud ecosystem cannot be entirely attributed to the players. Licensing - or the lack thereof - played its part, as did the technology itself. For many would-be cloud providers, F/OSS is simply the best tool for the job.
Greg Ness is extending and enlarging his theme regarding the demands that next generation datacenters and cloud computing make on infrastructure... and particularly network infrastructure. Notwithstanding the fact that he's now employed by a vendor of appliances and technologies that offers network services (like DNS, DHCP, IPAM, RADIUS, ...), the theme has merit -- it's not just a salespitch, folks. So, while I might argue with his analysis of VMware's fortunes, the basic message ... new approaches to infrastructure for next generation IT ... is dead on, and with it the requisite new approaches to infrastructure management.
... While many pundits have their heads in the clouds proclaiming the next big thing, there are a few issues that need to be resolved first. And those issues promise to fuel new demand for new types of networking solutions.
These new demands of scale and complexity and availability were beyond the wildest dreams of the creators of the core network services that support today’s increasingly strained network infrastructure. Many of these services, like DNS and DHCP are decades old. They were created in simpler days, usually in silos and with no concept of a need for interoperability between the protocols. Those days are now gone. DHCP servers, for example, now do dynamic DNS updates.
Very enjoyable, and very good, post from Amrit Williams on virtualization. He speaks to a number of popular notions (aka myths aka canards): - Virtualization reduces complexity. - Virtualization increases security. - Virtualization will not require specialization. - Virtualization will save you money today.
Now he does start to frost me a bit when he asks the musical question:
Want to guess how many start-ups will be knocking on your door to solve one or more of the above management issues?
Hey!! We're one of those start-ups!! Whatcha got against start-ups?!?
Thanks to VMware you can barely turn around today without someone using the V-word and with every aspect of the English language, and some from ancient Sumeria, now beginning with V it will only get worse.
Cisco announced some of the anticipated fruits of their partnership with VMware and their acquisition of Nuova (remember that Nuova's CTO Ed Bugnion was a VMware founder). As Ed says so well, server virtualization didn't break the applications but it certainly broke the infrastructure.
Michael Morris has a very succinct and informative post on Cisco's Nexus 1000V. The overview puts into perspective VN-link and includes a short interview with Doug Gourlay who mentions a few additional technology initiatives and goodies that weren't mentioned during the VMworld 2008 presentation.
Conforming to the axiom that it's easier to join 'em than fight 'em, Cisco launched its first software based network switch this week - the Nexus 1000V - as an integrated component of VMware's ESX platform.
This came as quite a surprise. Jeff Barr has written a short take on the announcement. But, what's as interesting as the fact that Oracle supports the use of their product on AWS is the larger issue of a big software company establishing a (potentially) workable licensing approach. Take heed.
We've been working with Oracle to bring a number of their products into the cloud. The first fruits of this work are now ready: cloud-compatible licensing, EC2 AMIs preloaded with a variety of Oracle products, support programs, backup to the cloud, and a cloud management portal.
As more and more enterprises take a look at the Amazon Web Services, they invariably ask about packaged software, particularly databases. With this announcement, AWS users now gain access to a commercial-grade, brand-name database, along with the necessary tools and middleware needed to build and host heavy duty enterprise applications in the Amazon cloud.
So, what's available?
Oracle_openworld The Oracle Database 11g, Oracle Fusion Middleware, and Oracle Enterprise Manager can now be licensed to run in the cloud on Amazon EC2. Customers can even use their existing software licenses with no additional license fees. Read more about cloud licensing here.
I should say a few words about licensing here because this question comes up all the time. The variability and flexibility of cloud-based licensing has perplexed users and vendors for some time now. Now that a large software vendor has made a clear statement of direction here, we should see more and more cloud-compatible licenses before too long.
Tom Bittman of Gartner has recently started blogging on cloud computing and virtualization. In a post made after the opening gun at VMworld 2008, he comments on two strategic shifts evident in the VMware story: infrastructure management (which he characterizes as throwing down the gauntlet with IBM, HP and MSFT) and cloud computing.
What interested me in the post are some of the presuppositions and his conclusions: (a) it's inevitable that the datacenter becomes a virtualized (b) in becoming virtualized, the virtual machine environment (in this case VDC OS) becomes the natural locus of end-to-end datacenter infrastructure management (c) by adding service governance to the mix, one has a management system that competes directly with adaptive, utility computing management strategies promoted by IBM, HP and Microsoft
While this analysis of VMware's strategy makes sense on its face, it also seems to couch the competition in terms of failed or stalled initiatives at (some of) the competitors. Bittman alludes to this in his commentary. For some reason, when thinking about datacenter operation, administration and management, I would have been more likely to set the competition as being between VMware (and its hoped-for coterie of infrastructure management partners) and the Big 4 (and Little 4) systems management providers.
The point worth noting: we need a more thorough discussion and definition of datacenter service governance (to use Gartner's terminology). This becomes critical, for example, when considering the discussion of VMware and virtsec and even more so when reading Hoff's consideration of network issues in the virtualized datacenter. Then, we'll be able to have a better conversation about how systems management in the datacenter actually comes to pass, and how VMware will compete with the Bigs.
VMware includes in their concept what Gartner calls a service governor, which adds policy-based management on top of a meta OS. Combined, these two create what Gartner calls a real-time infrastructure. The service governor is the real challenge for VMware, which is one reason they haven’t called it out.
What is interesting is that VMware is finally describing a larger strategy that is completely competitive with IBM (remember the On Demand Operating Environment?), HP (Adaptive Infrastructure) and Microsoft (Dynamic IT). The strategy is credible, but there are many, many gaps that need to be filled. In particular, while VMware is strong in virtualization, they are very weak in service management. Regardless, it will be difficult for IBM and HP to miss the competitive threat (which, of course, they should have seen starting in 2001). This is the only natural evolution for VMware, but the road is littered with challenges.
Over at my day job, we’re conducting a quick little survey to help us get a better handle on the configuration problems that people are running into setting up and managing a virtualized datacenter. We are looking for feedback on the challenges that frustrate administrators in a virtualized world, and then some specifics on areas that may be of particular pain such as IP management, network configuration, etc.
The survey should take < 5 minutes, so if you’re reasonably technical, have played with or administered virtualization (citrix, vmware, microsoft, KVM, whatever), please help us out and take the survey. Feel free to pass the link to anyone else.
I’ll be posting the full results of the survey (barring personal information) next month, no filtering or editing. If I get more than 50 useful results, I promise to even post the raw XLS for anyone else (competition or otherwise) to use as well.
Had an insanelyinterestingtwitterconversation this afternoon with John Mark Walker and Byron Servies. We’re in the process of hiring a few people, and I’ve had some sub-optimal candidates so far. I have the nasty habit of tweeting some snarky thoughts after my phone screens. Today, I had a candidate tell me that he wants to work 20 hours/week max in front of a computer. When pressed on what else he wanted to do, he kept talking around the topic, saying “I’m 3* now, I can’t work 60 hours/week”. He was digging for all sorts of deep questions that actually just made me uncomfortable.
This rubs me the wrong way for many many many reasons, though not the one you’re thinking of. I hope this doesn’t piss off any potential VCs, but I don’t think we’re actually doing 60 hour weeks here. Keep in mind that 60 hours is 9am - 6pm 7 days week, or 9am - 9pm M-F. I’m probably doing about 9am - 6:30pm, plus odds and ends on the weekend right now, averaging 50 hours. I have NO doubt that there will be 60, 70 hour weeks or more when needed, but the goal is to keep that to a minimum for everyone.
What got me going with this guy was the message behind it. No one wants to work insane hours. We all get that. Work life balance is one of the hardest things we all face - and I don’t have kids yet, so I have no real idea of the challenges. I really don’t think we’ve hit on the right balance in our culture, and I freak out a bit whenever I think about the future trends that will just make this worse. All that said, saying you don’t want to work long hours is just totally counter productive. Yes, as Byron points out, there are still many many managers out there that confuse face time with productivity. But saying you don’t want to work 60 hours/week doesn’t get at that. Nor does it get to the work/life balance. It’s a negative statement, it’s what you don’t want to do.
My advice here, is to focus on the positive, outline what you do want, and test what you don’t.
Focus on the positive
Instead of saying what you don’t want to do, tell me how you’ve kicked ass in flexible environments in the past. Or how you’re excited to make the team so productive that no one ever needs to work miserable hours. Or about the time you managed to outperform some teammate 4:1 and finished in 10 hours what he did in a week. Telling me the negative just makes you a prima donna.
Outline what you want
It’s fine to tell me you’re concerned about work/life balance. Tell me you are looking for a job that respects your family, and allows you to spend the time you need with them. Tell me you love to travel, but that for now you’re looking for a job here locally. But remember, I’m trying to fill a job, so make sure it comes back to how it’s going to help me fill the position.
Test what you don’t want
Blah blah, after all the above, what you really want to know is will you be here till 10pm every day. Guess what, no matter what you ask, you’ll never know. I may lie. I may forget. Hell, I may say yes cause I think that’s macho, even though I go home every day @ 4:30 to catch my talkies. Just your asking makes me cautious and nervous about you. So don’t ask. If you’re about to commit 2000+ hours of your life to a company, how about taking a few extra hours to drive by the office during times you hope people are at home. See how many cars are in the lot. Email the hiring manager a thank you note at a strange time, and see how quickly you get a response.
Remember, at the end of the day, you always can say no. Use the interview to sell yourself, do your research independently. Trust, but verify.
Although it took some time, I’ve decided on the tools. As the title says, I’m going with Atlassian’s suite, plus a really nifty plugin I’ve come across called Greenhopper. Here’s the overview:
Create user stories in Jira.
Create wiki page for each release (not iteration, but release). Link up to Jira dashboard. Provide details, necessary information, any supporting stuff available.
Planning poker to estimate user story size.
Prioritize user stories with drag and drop via Greenhopper.
Drag user stories into iteration releases via Greenhopper.
Create sub-tasks for each user story to track hours.
Drag tasks on planning board for open, in progress and done via Greenhopper.
Use Greenhopper to generate burn down, velocity, and other fun graphs.
The beauty of greenhopper is it provides a trivial and simple interface into a very powerful bug tool. You get all the benefits of a shared collaboration tool, without the pain. Greenhopper lets you edit directly in place, use drag and drop, and basically make it all so nice and easy.
Now, all that said, there are times when paper is just the way to go. We currently have ~85 user stories, and trying to figure out the ranking really is easier when you can have paper in front of you. Luckily, it’s easy to do both! By exporting the issues into excel, and then using mail merge from word, finally printing to Avery postcards, I’ve got a great way to create cards. You can use my attached user-story-cards word file if you want to create your own.
Ah, the corporate overview. The staple of every web site, and the key piece of many presentations.
We’re in the proess of putting such collateral together here at Replicate. Not wanting to reinvent the wheel, I consulted the oracle for any keen insight or best practices. After 45 minutes of searching, I’m surprised to find nothing on what makes a good/bad overview, or thoughts in general. I did find some amazingly, painfully, searingly bad examples (I’ll be kind, and not link, but a quick google search for “corporate overview” will send you crying for some good clean smut).
Obviously, the goal is to convey answers to the standard 6 questions as applicable - Who, what, when, where, how, why. I think we’ve all been trained a bit too well on these in that order. Is who really the most important part of any pitch? Unless it’s to your mom, I don’t think so. Keeping in mind who we’re trying to convey this information too, most people who haven’t heard of us don’t want to know about the company, the want to know why they should even care in the first place. To steal from Jerry Weissman, WIIFY (what’s in it for you)? The WIIFY will vary depending on your target audience. In our case, we have a few, all rather obvious:
The customer: “If you buy our product, you’ll solve a ton of problems and get to go home and spend time with your family”.
The partner: “If you work with us, you will keep and gain many happy customers”
The investor: “If you invest in our company, you’ll get a great return on your money”
In all these cases, talking about our company seems like the last step, not the first. Identify their problem, talk about solution, then explain why you’re the one to solve it. I’d take those favorite six questions and reorder them:
Why: What’s the market problem
When: I’m gonna care about this when?
What: What is the solution?
How: How will you help solve it?
Where: where can I find the solution?
Who: And who are you to actually do it?
Any other thoughts on outlines or best practices? What’s key in a good corporate overview doc or presentation?
I’ve concluded the tools analysis that I kicked off a few weeks ago. What initially began as a look into requirements management tools quickly expanded into the entire product development lifecycle. Tools can not make up for broken process, so that’s the place to start. As I looked back at what’s worked and not in the past, I realized that the MOST critical success factor has always been alignment. If engineering, product management, sales and the customer aren’t all aligned on what, when, how, and why, we create massive headaches for ourselves. To ensure alignment, we need one source of truth, one place to track everything that’s going on. If we need to keep two system in sync, we lose. If everyone in the company can’t see everything, we lose. If everyone isn’t bought into the process, we lose.
In order to get everyone on the same page, we need to look at the toughest participant. I’m sure this won’t come as a surprise to anyone, but usually that’s engineering. Not only are they the largest team, they also have the strongest opinions. Starting from that vantage, I realized the plan of record has to be kept in a bug database. It’s the only thing engineering will update. In my past lives I’ve many times tried to keep a separate list, one focused on features, and engineering’s DB focused on bugs. But what happens when you classify a bug as a feature enhancement? It’s such a grey area, do you start replicating and moving data? Duplicating it? Either way, it’s a disaster. I’ve gotten into massive fights over this before, and even if I was “right”, is it worth fighting that over and over?Yes, tools like Rally and Accept can interop with Bugzilla or Jira. But if I need to have the bug DB anyway, is the extra effort and cost worth while? For such a small shop as ours, I’m not sure the answer is yes. I’ve used Accept for a few years, and have loved the tool. For traceability and analysis, it’s the perfect tool. I think I may have let those aspects get in the way of focusing on what the requirements themselves are, and how we get alignment in the groups. Frankly, all the tools I looked at in the previous post just seem like total overkill for any <50 (and maybe even <300) person company.
Now clearly, the bug database by itself isn’t sufficient. You need a place for more unstructured data, for planning, notes, etc. For marketing data, sales plans, and just the rest of the company. Obviously, a wiki is the best place to put this. This leaves us needing a great bug DB, and a great Wiki. In a stroke, I’ve thrown out almost all the tools I was looking at before: Accept, Rally, Feature Plan, VersionOne. They all are… complex. I’m sure for a large company, they’d work great, and have their advantages, but for a startup looking at Scrum and frankly trying to save some money, I just couldn’t see the value. So that leaves us with only four options that I could see out there:
Fogbugz (with or without external wiki)
Twiki + Bugzilla
Trac
Confluence + Jira
Fogbugz I quickly ruled out. Although it’s got a great UI, and I hear good things about it, it’s also just not the right fit. Some of the additional features were lackluster (the wiki is atrocious, the forum useless compared to phpBB), plus it’s very general (good), and needs to be shoehorned to fit Scrum or other agile processes (bad).
The next three was a much tougher decision. Trac looks great. For an internal focused project, or an open source project, it’s probably the one I’d use. However, we have need for both private and external content, tighter security controls, and some more advanced commercial features that they don’t seem to cover. I REALLY liked Trac, but I didn’t see it fitting into our commercial environment.
Now it’s really hard. I LOVE twiki. Better than Confluence. I like Jira better than Bugzilla. Ultimately, the support & integration that Atlassian is providing, plus the larger suite of products such as Fisheye, Bamboo, etc all pushed me over the edge. Maybe it’s just a dream, but the thought of having a view into CI data, test status, check-ins, all tied and traceable, makes me a tingle.
From a cost perspective, for a small team like ours we’re looking at a $3,600 investment for Jira and Confluence, plus $1,800 for maintenance. We’ll put them in VMs on existing HW, running Fedora. The incremental admin costs for two new machines should be minimal. I’m figuring if we buy into their whole stack, we may wind up spending 10K with them. On the one hand, that ain’t chump change, especially when we could be getting it for “free” with Twiki/Bugzilla or even Trac for internal + Twiki for external wiki needs. I’m counting on the support, the features, and the overall new development will all make this pay off.
I know many other people out there on the internets have gone through similar evaluations. Which way did you go? 6 months or a year later, are you happy about it?
BTW, the photos are from a recent drive I did. They are there just for decoration.
If you follow me on twitter, you may have noticed that I slipped in my resignation from Sun. Today, March 14th, is my last day.
I’m embarking on the riskiest of all options - a tiny new startup. We’re looking at how we can address some issues that arise with the use of virtualization and network management. I’ll be doing product mangement, plus whatever it takes (markeing, cooking, cleaning) to make sure we’re a huge success, and maximize the reward side. [If I've 10x my risk, but also 10x the potential reward, I guess my risk/reward ratio remains constant? I guess I've probably 10x the risk for <10x the reward.]
So why jump now? For the first time in my career, I’m not running away from an old job. The Sun xVM team, and especially my boss, are doing an amazing job. They are building a great team and a great product. I’ve even managed to recruit two new people to the team in the past week, in spite of my departure! I have no doubt that I’ll be talking with the Sun xVM team shortly on how we can work together.
Proving just how random our lives truly are, I found this job through one of those serendipitous moments. A colleague of my wife asked if she knew anyone for a certain position at this other small company, just in passing, and a month later I’m resigning! The team, the opportunity, and frankly, the risk, were all to good to pass up. I’m excited to be scared sh-tless and I’m excited to make something from nothing.
Rich, Rich, and Ken - I can’t wait to join you and the rest of the team on this adventure.
