The Hindu Business Line : From RIM to centre-stage

Remote infrastructure management is set to rise from $7 billion now to $25 billion by 2013. And India has a chance to capture a 50-55 per cent slice. How can it be done, and what are the challenges? eWorld captures a lively round-table as it unfolds.

K. Bharat Kumar

A tiny seed produces a large tree. What was a 4,000-manpower industry five years ago, is now filled by 40,000 people. In the next five years, it could well reach a headcount of some 3.7 lakh. In revenues, it could touch $13 billion in 2013 from a current $3.2 billion. Welcome to the remote infrastructure management (RIM) industry in India. RIM refers to the clutch of vendors who provide maintenance support for hardware and network infrastructure to clients in the Western w orld, from an ‘offshore’ location such as India.

A fortnight ago, Nasscom, the apex body for the IT and BPO industry, along with McKinsey & Co, announced that it had come out with a report, bringing out the numbers indicated above. In summary, what it said was that the global addressable market for RIM is $104 billion. And India has a chance to capture 50-55 per cent market share in that segment. In this context, Business Line and HCL Technologies jointly organised a round-table to discuss the prospects for the industry. Vivek Pandit, Principal Analyst, McKinsey, Ross Tisnovsky, VP-ITO Research, Everest Research Institute, Anant Gupta, President, HCL Infrastructure Services Division, and Carlos Cabrera, CIO, Exide, joined us in the chat. Excerpts from the very lively discussion:

Given the visibility of RIM in the Indian media, one would think the sector had reached saturation. But the numbers obviously indicate potential.

Ross Tisnovsky

Tisnovsky: Indian companies started work here based on their application development and maintenance (ADM) portfolios. These relationships are getting exhausted. A lot of players cannot grow anymore based on these relationships. This market has reached saturation.

For the rest of the market, we are extremely far away from saturation. If you look at the mid-sized market, Indian companies have better penetration. For the large clients, it’s only the beginning. Indian vendors are in the ‘penetrate-and-radiate-mode’ for the large buyers. The deal sizes there are very small. Deals in the mid-sized market are larger than deals among large buyers, (defined as $10 billion in revenues).

Vivek, what gives you the confidence that there is such growth possible? Give us some insight into what went into the making of your report.

Vivek Pandit

Pandit: As a result of declining hardware prices and increasing virtualisation, labour is sticking out as the single largest contributor to costs, inside the infrastructure space. Labour has become Moore’s laws’ latest victim. Given the plummeting equipment costs and consolidation and standardisation taking place inside organisations, suddenly labour sticks out. What an offshore vendor could do to create value for customers was limited seven years ago. Now, it’s different.

Carlos Cabrera

Cabrera: The key to the outsourcing decision is whether the skills and talent to implement what we wanted and to support it were available. Once you decide to take it off your premises, it does not matter whether you outsource. Anyway it’s going beyond your walls. Whether it is in Mumbai or Minnesota is irrelevant. For us, the trend I see is to offshore even more than we currently do, with HCL Technologies. We are facing a lot of pressure to reduce our SG&A, our (other) costs and if we want to do that we have to be flexible as to where we go. As a medium-sized company, five years ago, I would not have thought of it. Now, it has become practical for us.

With all the tools and algorithms and process maturity that you guys have brought in, are we anywhere close to offering desktop services remotely at $1 per month per desktop?

Tisnovsky: There seems to be a sudden explosion of innovation in the desktop management area which has not happened for a long time. I don’t just mean the ability to remotely login to the desktop and perform some trouble shooting there. Intel’s recent announcement of the vPro platform is interesting, because it allows you to log onto a machine that is practically switched off. It has humongous implications for RIM, because you can log on remote into a machine that is not even switched on. If this works, and I will allow myself a degree of scepticism, but if this works, it is a major step up on offshoreability of desktop support. At $1 I don’t know, but at least Anant can deliver it without people onsite.

Anant Gupta

Gupta: If you dissect what that service means — we call it end-user computing service — first comes the basic level help desk; second, is the advanced level with some application support; third is software distribution layer; fourth comes asset management; fifth is application packaging or imaging; and sixth is the hardware break-fix onsite. Of these, pretty much all of the first five can be done offshore. The last piece is a challenge. In India, we manage financial networks across 260 towns… towns could take as many as 10-12 hours to get there. We have learnt to provide the right measure of support for a hardware failure in this diverse geography.

With these remote diagnostic tools to help you can reduce onsite call rates by 60 per cent. Benchmark cost savings are about 40 per cent of the total cost of $130, again, on average.

Pandit: As we see it, in terms of the opportunity, desktop support ranks lower down: towers, midrange, network, mainframe, help desk, and then desktop. Help desk is like Anant said it — you seek ways to improve first call resolution rates and to reduce handling times. These are the main drivers of cost and hence of value. From an offshore point of view, the ability of vendors to do this is high.

The single biggest driver of cost on desktop is software, not labour. Hardware prices of desktops have reduced considerably. Prices of the operating system and the applications have to come down. Offshore vendors’ ability to manage that cost is non-existent. Second is the complexity of desktop — some are superloaded and the others aren’t. So it’s difficult to give an average…

Cabrera: We have machines still running Win 95 on them!

Pandit: A more interesting comparison would be the cost of managing a mailbox. That is an easier benchmark to set. That is a little more standard — can be priced as per amount of storage, etc.

Gupta: There are some benchmarks. At the low end, globally, outside the lease of the desktop and the licence costs, it’s about $90-95 on average and at the high end it’s about $160-170 per desktop per month. It assumes a certain ratio of end user calls to the support desk. Pretty much, you can do a lot with the $90-95 itself.

Ross, you had earlier said that Indian companies are not aggressive enough in terms of acquiring asset-heavy resources. Any views on the way the market has panned out since then?

Tisnovsky: You need to define aggression. We get comments from multiple traditional suppliers. They read ours and others’ reports. They tell us that they keep hearing of the emergence of remote offerings but that they don’t see them in those deals. What is often missed is the Indians’ penetrate and radiate strategy. Indian companies often grow their renewals. A deal is first signed and it’s a tiny deal. When it comes to renewal, there is a huge uplift in score. The offshoring industry is shy about talking on deals. They are not announced. They go nice and smooth in the background. This growth through renewals is not apparent in the market. We try and pick it up through our requests for information (RFIs) and analyses of deals, if anything it’s increasing. We can talk about this passive or hidden aggression.

Gupta: I have a different opinion. They are not necessarily small and home grown. About 12 or 13 deals are to be decided this quarter and we are fighting about 10 of them with the big boys. The success rate between insourced and outsourced is extremely high. Renewal, where there is a large player, is seeing lesser penetration at this point. Be it in North America or Europe, we see a completely different trend — it’s RIM vs total outsourcing. It’s not between Indian and global players, it’s between the business models. Once they have homed in on the model, it’s common sense to decide what players to deal with.

Pandit: As a result of the work we did through Nasscom, we saw three things happening: the number of request for proposals (RFPs) is up significantly; the win rate of those RFPs is up significantly, so they are winning a larger portion of more RFPs; the competitors to these RFPs are increasingly the top six.

Without question, in some of these cases, offshore vendors are being invited to bid. You would see two-three offshore vendors among two-three traditional vendors. That is the seriousness with which customers are thinking of in regard to Indian vendors.

In terms of acquiring assets, are Indian vendors a bit more aggressive than you earlier thought them to be?

Tisnovsky: A lot more now. Wipro’s acquisition of Infocrossing is an example. They will be taking on assets; not necessarily transferring assets to their data centres, from someone else’s books, but delivering their services from their servers, which is essentially the same thing. So there seems to be a change. The way we see it — labour arbitrage alone is not a value proposition long term, although the ratio of assets in the expense pie is declining, pretty significant, but doing something about assets will become critical for Indian suppliers.

When everything moves offshore, do we have the infrastructure pieces, power being one, to manage that?

Gupta: The physical assets don’t move. They stay in onsite locations. Management of those assets is taken care of offshore.

Even if you have to think of transferring those….

Pandit: Four or five years ago, we analysed the cost of moving an entire data centre to a low-cost location. At the time, there was a 15 per cent increase in costs, a lot of that came from the cost of licences for software and cost of procuring hardware, which were higher in emerging markets. Leave alone managing, but just setting this up physically was costlier.

We ran the same study 18 months ago. On a like-to-like basis, removing power from the equation, the move saved you about 5 per cent. Not enough to frankly justify the move. But now, whether you are now moving from New York to Washington and the labour is in New Jersey or in India matters to an extent but not like it used to.

Second, infrastructure does move to offshore locations such as Singapore and Malaysia. Those are largely disaster recovery centres (DRCs) or redundancy centres. There is a strong emerging proposition in South-East Asia where clients want an alternate site. India won’t figure in that list since our ability to support such moves is suspect.

Unfortunately, our infrastructure makes us uncompetitive. Our figure of 50-55 per cent of market share would be higher but for our infrastructure.

Should a slowdown/recession hit the US and hence RIM, would the challenges of a RIM provider be any different compared to an ADM provider? A small client is already uncertain about the macro economic conditions. Why would he add another piece of uncertainty by going offshore?

Gupta: RIM comes from non-discretionary spend of the budget. You need not run it so that you super excel but you need to sustain and run it. The imperative to look at RIM is even more compelling if the discretionary spend is curtailed. There may be a need to develop a business critical application, or for a technology refresh, so you then offshore IM more and use the cash for new investments. That is, you get cash from the run side of business and put it into the build side of the business.

Tisnovsky: In application development, it is easy to stop everything and cut costs. In infrastructure it’s difficult. That’s why a recession is bad news for ADM. In infrastructure management, it’s not necessarily bad news for suppliers who prepare themselves for it. If you want to consolidate data centres, you hire someone to do it. A lot more infrastructure outsourcing may happen in the case of a recession.

Cabrera: When we decided to move offshore, we had lack of funds. RIM helped us reduce the cost of running things. We took those savings and used it for a technology refresh. Savings from offshoring helped build more infrastructure.

What are the challenges on the manpower front?

Pandit: Challenges similar to those faced in ADM. In RIM, a lot of the roles, the skill-sets required, are higher than those in BPO and some aspects of ADM. While they can be trained, training takes more time. The tools you need to invest in and the training on tools are more complex. However, RIM, relative to ADM and BPO for a lot of people in the market, is a lot of more attractive. It’s a lot closer to technology. Those who did go through engineering schools and not just vocational training, tend to move into technology environments. The nature of relationship in RIM is long term, so you are not churning contracts over three-six months.

What is the attrition in RIM? IT services averages between 11 and 16 per cent.

Gupta: RIM is about 14 per cent. Outside of the helpdesk section, it is lower. People here need to be a bit more different from ADM. In a Java project, the front 10 guys need to be senior while the 100 back room guys could be freshers for all you know. There are good checks and balances before the system hits the client. In RIM, it’s different. The maturity levels required, training, etc, in RIM is high.

Attractiveness is there since it’s a lot more close to technology. Engineers typically want scale and complexity in the servers they manage. If you are a UNIX guy in your fourth year, you want to manage 300 servers. They had to move around the globe looking for opportunities. Now it’s different, technology and scale have advanced, there is good money, careers, lifestyles. We see a lot of technology moving back here.

From $3 billion now to $13 billion in 2013, what should the Indian RIM industry do differently?

Pandit: First and foremost talent. It does not happen on its own. If you look at courses offered in computer sciences in US, a number of them are based on MIS, are infrastructure related… there is a sort of vocational training in polytechnics as well. We don’t have that in India.

We are constrained in part by regulation as to what extent government colleges and universities’ curriculum can be influenced by the private sector. Over time, just as CS programs adapted to the needs of the industry, we have to introduce a number of courses relevant to RIM.

Cabrera: Universities in the US are complaining about the fact that students are not going into CS, as they used to. They are not going into engineering to begin with. It’s because a lot of people have read that computer jobs are going away. So the thought is, why go into a field that is going away? It’s almost creating a need to offshore since you are not getting that internally. I myself ensured that my two youngest daughters went into vocations where it is difficult to outsource — one is a veterinarian and the other is a lawyer (Smiles).

Gupta: Five years ago, we used to get an engineer on board, train him on Cisco technology, get him CCNA certified, and then bring him on. We spent 12-15 months on this. We see that now happening of its own accord at the university program level. Engineers now realise that a CCNA means better opportunities. So, they sponsor themselves.

Pandit: There’s regulation right now that regulates working hours. Apply the queuing theory to create efficiencies in your model by shifting utilisation. In RIM, you’d be better off working three days on and getting two days off. But the Indian labour laws disallow working hours beyond a certain number per day. Weekly hours are fixed, but there is no flexibility around the daily hours. Encryption laws also pose a problem. It creates further onus on the vendor to manage the relationship with the government.

From the perception point of view, there’s something perverse about saying “offshore infrastructure to India”.

Infrastructure is not something that comes top of mind when you mention India. The amount of rhetoric on India managing someone else’s infrastructure is almost ironic.

The industry talks of increasing the 40,000-manpower strength to about 3.72 lakh in 2013. Is there a management constraint in terms of the ability to manage this size?

Pandit: We have two Indian IT vendors who have crossed a lakh in headcount. The ability to manage employees numbering so many is something that has been shown and proven. The management of IT talent at that level is happening in India. On the question of RIM, this is not purely a full-time equivalent (FTE)-based model. That is, while the industry is projecting 3.72 lakh, etc, it would depend on the models deployed. You could have pricing models that are device-based. So, you’d pay based on the number of servers being managed, as opposed to paying for a certain number of FTEs. When you pay based on numbers of servers, a vendor can improve productivity so servers managed per FTE continues to rise.

Cabrera: In our case, one of the things that was different from our previous provider (who was also a global player), it was server-based and that created a lot of problems for us. It sort of limited us.

Every time we needed a new server, our ongoing maintenance continued to rise. Whereas in the current arrangement, it is FTE based — we just indicate the number of FTEs required for the number of servers we first start with.

But as you add servers, you only need to add FTEs that are dedicated when the capacity of FTEs exceeds their ability to manage the servers.

It was an add-on every time we wanted something new.

Ross: I think people often miss another aspect of management complexity, that at the buyer-end. We’ve been talking about suppliers’ ability to manage more people — hire and retain employees. Can clients manage complexity of multi sourcing relationships, which is likely when you go with India suppliers? And could they manage an offshore company, which requires a certain degree of collaboration. Carlos makes it appear very simple, the way he talks about it. That’s possibly because he is at the top X per cent in terms of managing vendors. If you look at the buyers’ community, many might not have enough sophistication to consider these complex environments. One of the things suppliers could do is to transfer this complexity to themselves, but then they would need to build levels of sophistication for this.

Gupta: To add to Ross’ point… Exide is in a different league, of course. You need management skills. But, you also need the mandate to be the aggregator in-house. A lot of companies, that had previously outsourced, got rid of internal IT management skills and found it difficult to be the aggregator of multisourcing agreements. They need to build that capability and take advantage of the RIM and the multisourcing models.

Cabrera: Even though we have outsourced a lot of our applications development and infrastructure maintenance, we have subject matter experts. Someone is going to oversee how messaging is done or how security is done. We are not going to give away all that knowledge. I am still going to make people at Exide responsible for driving change.

So, how do you manage complexity, especially when the vendors number so many?

Cabrera: Even within Exide, many people wanted only one vendor. But I feel (differently). You aren’t going to have 10, but I would be uncomfortable having one. But three-four providers are good, if you pick them well.

Gupta: Just to clarify, across the outsourcing spectrum, you are going to have expert vendors in the areas of RIM, BPO, ADM, network, the telecom players…. Good eco system based on best of breed.

So, given all this good work done these past years, there is less of a threat from alternate geographies such as China?

Pandit: No. India benefit is not so much the differences among ADM, BPO and RIM. It’s the similarities of challenges that you face across. Meeting the challenge of creating talent that might have gone to other industries but making them available for this industry; that’s something Indian companies have cut their teeth on before.

Tisnovsky: I have a question to which we don’t have an answer. Only young Indian professionals do have an answer to that. There’s a fundamental disadvantage that India has — different time zone. In ADM, non-stop 24 hour production. So we can do things off hours. It’s not going to happen here in RIM. By definition it’s real time. Whatever you do, tools, etc. Someone is not sleeping here.

Indian professionals are resilient and willing to do this. But for how long? Is there an alternative? You can go to ADM, or even real-estate and sleep during the night.

Ross, Everest had some RIM industry estimates for 2010. Do those tally with McKinsey’s?

Tisnovksy: Ours was $8 billion by 2010. McKinsey’s is $8.2-9 billion. It’s close.

Cabrera: I have a question. How is Software as a Service (SaaS) going to affect the way you support clients? All the applications are somewhere in the cloud and the need for infrastructure for a client becomes less.

Pandit: If you look at salesforce.com — a lot of buyers are mid-sized or smaller companies. Some of those companies were not even using those applications. Some who were, are now saying ‘I don’t have to host them but get the service from someone else.’

It would drive consolidation among those providing hosting services. A lot of smaller and mid market hardware footprint would go into a single location.

Another result is higher utilisation. You are potentially sharing racks, not cages or facilities. It has a dampening effect on the amount of hardware getting procured. But the amount of service provided because of the hardware spend and getting scale out of it would give rise to even more opportunity for remote management.

Tisnovsky: I agree. There is an interesting untapped opportunity for SaaS. It’s a huge opportunity and we don’t see anybody building enough capabilities to do that.

When you outsource/offshore, RIM, how about security administration. Some part of it is also offshored, now. But if you are giving it to the same guys, it’s like the fox guarding the chicken house...?

Cabrera: Security is one of the tasks we have outsourced. We do have someone internally who is going to manage that. We do work for the government so we did have concerns regarding security. We segmented portions of that so that critical information stayed domestically within Exide.

Tisnovsky: Could you please elaborate on the aspects of security — perimeter security, communication security, encryption...?

Cabrera: The management of active directory or addition of users, user security and also the perimeter security, firewall (outsourced to the network provider, BT globally, as we converted from Frame Relay to MPLS network.)

Pandit: It’s the management of security that is being outsourced. The policy, supervision, etc, are still in-house.

Gupta: The infrastructure security moves, but security policy and planning remains in-house — that is the key part.

Tisnovsky: Essentially, you outsource the key strokes of what someone has to do. You outsource the process. The policy, who gets access to what, is in-house.

Cabrera: It has to be. Else we wouldn’t pass the SoX test (smiles) (Sox refers to the Sarbanes Oxley Act, which lays down the law for regulatory compliance)

Gupta: We have seen highly regulated industries outsource security quite a bit. Banks and Pharma companies have done it, despite being on the top end of concern curve. Nothing scary because it’s the operations, while the policy is retained.

bharatk@thehindu.co.in

More Stories on : Interview | Insight | Software

Article E-Mail :: Comment :: Syndication :: Printer Friendly Page

Stories in this Section
Buzz at Barcelona

For healthy clicks

System Restore in Windows XP

From RIM to centre-stage

Scoring with service delivery model

Quiz

Sponsors, guard against surprise attacks

Cartoon

Hear, hear

A slice of Apple

Sound of clarity